<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Paul Maltby on Medium]]></title>
        <description><![CDATA[Stories by Paul Maltby on Medium]]></description>
        <link>https://medium.com/@maltbyps?source=rss-d3a206a02212------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*PPvtXWnrQIp_emtGlgnQSg.jpeg</url>
            <title>Stories by Paul Maltby on Medium</title>
            <link>https://medium.com/@maltbyps?source=rss-d3a206a02212------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 01:56:32 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@maltbyps/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Reasons to be optimistic about DSIT and the new digital centre of government]]></title>
            <link>https://medium.com/@maltbyps/reasons-to-be-optimistic-about-dsit-and-the-new-digital-centre-of-government-f6813b8e8043?source=rss-d3a206a02212------2</link>
            <guid isPermaLink="false">https://medium.com/p/f6813b8e8043</guid>
            <category><![CDATA[digital-government]]></category>
            <category><![CDATA[dsit]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[gds]]></category>
            <category><![CDATA[policy-making]]></category>
            <dc:creator><![CDATA[Paul Maltby]]></dc:creator>
            <pubDate>Wed, 24 Jul 2024 16:54:06 GMT</pubDate>
            <atom:updated>2024-07-24T16:57:50.818Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*L4ZNqOgZOYQL3EtOaz6BRw.png" /><figcaption>DSIT HQ, 100 Parliament Street, London</figcaption></figure><p>For those of us interested in digital and data side of public services in the UK, one of the early consequential moves of the new government was the <a href="https://www.gov.uk/government/news/dsit-bolstered-to-better-serve-the-british-public-through-science-and-technology">move of GDS, CDDO and I. AI into the Department of Science Innovation and Technology (DSIT)</a>, under Secretary of State Peter Kyle MP and digital and AI minister Feryal Clark MP. Over the last couple of weeks I’ve had a few conversations with civil servants who have been scratching their heads at this move, some wondering why those teams have been ‘downgraded’ from the Cabinet Office into a new, small, regular Whitehall department. Although it is still early days, personally I’m optimistic about the change and I thought it could be useful to set out why.</p><p><strong>First, never bet against a ministerial team with energy, political backing and competence.</strong></p><p>The first reason I feel hopeful about the new DSIT team is that in my experience political weight and focus beats institutional hierarchy almost every time, and especially so in the early days of a new government when energy is in plentiful supply. The highs and, especially, the lows of the Government Digital Service over the last decade suggest that a Cabinet Office email address alone is insufficient to imprint effective and strong leadership over Whitehall. Francis Maude exerted such a strong influence in the early days of the Government Digital Service through his political skill, personal proximity to the Prime Minister, his ministerial longevity, his grip over administrative levers (such as spend controls) and his boldness in appointing a team of mainly external specialist hires with a track record of delivering the type of change that the traditional civil service had struggled with. There is every sign that Peter Kyle and his team have the zeal required to restore this ministerial focus for a second decade of digital public service reform, a direction that has been sorely lacking in recent years.</p><p>But to make the new arrangements work the DSIT team will need to use all their statecraft to get things off to a good start. As<a href="https://www.notion.so/gds-dsit-blog-18-7-24-ba47faee01464027bb943cdc4ccfac34?pvs=21"> Geoff Mulgan wisely notes,</a> Whitehall is good at ignoring cross-cutting programmes if they are only based in another department rather than in the centre, and the new arrangements do introduce a degree of fragility that will require excellent relationships with the PM, No10 and the Cabinet Office to mitigate the risks. But, we have seen it done before — I worked for the super impressive Louise Casey in the cross-cutting<a href="https://en.wikipedia.org/wiki/Respect_agenda"> Respect Taskforce</a> that was based out of the Home Office in the mid noughties, and her ability to use a mixture of political heft, Whitehall governance and external front-line expertise to make a difference across government and local government was hugely impactful — and same broad approach continued into her<a href="https://en.wikipedia.org/wiki/Troubled_Families"> Troubled Families</a> programme after the 2010 change of government.</p><p>So what can the DSIT team do to ensure their approach is a success? In my view they need to demonstrate that they have and are willing to use the harder levers of state (money, rules, appointments) to effect change across Whitehall, but they needs to choose the areas of intervention carefully. This is not a 2010 ‘<a href="https://www.gov.uk/government/publications/directgov-2010-and-beyond-revolution-not-evolution-a-report-by-martha-lane-fox">revolution not evolution</a>’ moment, when digital teams in Whitehall were either non-existent or captured by a failing approach; these days every government department and arms-length body has decent digital capability and the DSIT team will need to show these various effective and increasingly powerful teams scattered around our public services that they have something useful to offer from their central position.</p><p>There are multiple contenders for where they might focus, but a start will be to ensure a deep connection with the cross-cutting governance of the new ‘<a href="https://www.instituteforgovernment.org.uk/explainer/mission-driven-government-labour">mission-led</a>’ approach to Whitehall. Unless digital teams in Whitehall feel a strong and proactive pull into their department’s organising frenzy around the missions they risk being frozen out of the all important early hubbub of a new government, and this will then get reflected in funding at the forthcoming Spending Review. DSIT’s ability to quickly connect digital as a crosscutting enabler of the missions within the formal Whitehall governance processes and into the Spending Review will be an important signal of their influence, and that will be noticed and appreciated by senior leaders across government — in and outside the digital professions.</p><p>Another area where the new DSIT team can — and arguably already are — making a difference is on AI. The previous government picked up on the AI safety risks associated with frontier models, and<a href="https://www.aisi.gov.uk/"> AISI</a> is a small team with global impact. But outside this, the culture and practice of government digital and data teams have been too slow to understand and deploy the potential for AI within their core digital services, and in a world of endemic public service backlogs and eye-watering cost pressures the potential of AI can no longer be ignored, or seen as a surface level Ministerial plaything.</p><p>Setting out a vision for the importance of AI in digital services (busting backlogs and defeating drudge work!) and establishing an exemplar programme that allows departments to demonstrate what good looks like in practice (cf<a href="https://www.gov.uk/government/publications/gds-transformation-programme-2013-to-2015/digital-transformation-exemplar-services"> the original GDS examplars programme)</a> would be a great start. It then needs backing with money, access to top quality external capability, repeatable patterns and components, and expectations for how government teams should implement all this in practice through an updated service standard and spend controls.</p><p>One more area the new DSIT team might look at is to use horizontal levers of power alongside the traditional hierarchical government machinery. It has been an age since Jeremy Heimans and Henry Timms wrote about<a href="https://hbr.org/2014/12/understanding-new-power"> New Power</a>, the now near ubiquitous idea that in the internet age influence can be more powerful when distributed and harnessed rather than tightly controlled and managed from above; albeit Whitehall has been largely impervious to any competing narratives of power. Despite this there is a small army of reform-minded digital service workers across the state — in Whitehall, ALBs, local government, the NHS, Policing and beyond — who in recent years have had increasingly effective professional leaders in their own organisations, but the national context made it near impossible to maintain a groundswell of ambitious, collective cultural leadership at a whole-government level. Working with the grain of their own somewhat decentralised power, the DSIT team should aim to marshal this community of practice again, bottom up, learning the lessons from bold and energetic campaigning communities such as<a href="https://www.oneteamgov.uk/"> One Team Government</a>, who’s fabulous and<a href="https://www.linkedin.com/in/kit-collingwood-6674a091/"> energetic</a><a href="https://www.linkedin.com/in/jamesreeve/"> leaders</a> have since fled a Whitehall that struggled to make the most of their talents. One of the longest-lasting impacts of the original GDS is that their <a href="https://gds.blog.gov.uk/2016/06/23/good-culture-evolves-from-the-bottom-up/">posters and memes</a> are still found right across government — and that radical cultural leadership created an impact far beyond the reach of formal government processes. Whitehall’s<a href="https://en.wikipedia.org/wiki/Humphrey_Appleby"> Sir Humphreys</a> will find it harder to ignore an agenda set from a ‘lowly’ Whitehall department if their own teams are energised and feel backed by strong political and professional leadership from DSIT.</p><p><strong>Reconnecting the false divide between ‘digital’ for public services and the wider economy</strong></p><p>The second big reason for my optimism during this moment of transition is that the move of government’s internal digital and AI leadership to DSIT helps restore a missed opportunity from the last decade of otherwise mostly positive digital reform — for any government going all out for economic growth and productivity improvements we would expect to find a deep connection between digital competence in internal public services <em>and</em> in the government’s wider external curation of the economy and society. In recent years we have had plenty of the former, but only glimpses of the latter. The new shape of DSIT has the opportunity to change that for the better:</p><p><strong><em>a) Digital public service reforms that enable wider productivity boost: ‘government as a platform’</em></strong></p><p>One of the few data policies mentioned in the 2024 Labour manifesto was a commitment to a ‘National Data Library’ to support research. This is presumably a more open access version of the effective (in theory, harder to use in practice)<a href="https://integrateddataservice.gov.uk/"> Integrated Data Service</a> run by the Office for National Statistics, which is currently only available to government and accredited academic researchers. Where I work now, <a href="https://faculty.ai/">Faculty</a>, has previously worked with the Department for Education on a trial for an<a href="https://integrateddataservice.gov.uk/"> open version of the Longitudinal Educational Outcomes (LEO) dataset</a>, which used synthetic data techniques to protect this incredibly useful but deeply personal dataset so it could be used by more people in the research community, and there certainly is scope to better serve society and the economy by making data like this even more accessible by using the latest synthetic data and privacy preserving techniques.</p><p>But the potential for AI use cases based on government data goes far beyond the traditional and relatively slow-moving statistical-type data seen in the Integrated Data Service. Across government there are thousands of (digital) services that continually process data, and that data has a potential for driving more radical productivity improvements across our economy.</p><p>For example, in the Department for Levelling Up (now, thankfully, returned to the Ministry for Housing, Communities and Local Government after being de-sloganised) I set up a team that implemented reforms to the software that drives the public <a href="https://www.localdigital.gov.uk/digital-planning/">planning system</a>, and we did this in a way that ensured that the<a href="https://mhclgdigital.blog.gov.uk/2022/06/28/digital-planning-reform-an-overview/"> underlying rules and data could be externalised </a>so that our own teams but also PropTech companies and others could use use and trust that data, for example by building their own digital services on top of this data if they wanted. One of the planning system’s most fixable challenges is that its horribly analogue data is hard to find, use and trust. The MHCLG team’s developer hub<a href="https://www.planning.data.gov.uk/"> https://www.planning.data.gov.uk/</a> focusses on external software developers, akin to the sites we are all more familiar with such as<a href="http://developer.google.com/"> developer.google.com</a> or<a href="http://developer.spotify.com/"> developer.spotify.com</a> and countless others across the internet. In the UK government, however, such developer sites are few and far between — examples include HMRC’s<a href="https://www.tax.service.gov.uk/making-tax-digital-software"> Making Tax Digital</a> and DVLA’s<a href="https://developer-portal.driver-vehicle-licensing.api.gov.uk/availableapis.html#current-available-apis"> driving licence developer portal</a>.</p><p>For the new DSIT, aligning economic growth from the digital economy <em>and</em> the internal digital reform of public services, there is a huge opportunity to drive similar user-focussed developer hubs with useable APIs that point towards different sectors of the economy to drive digital productivity growth in areas of the economy from housing and planning, health and life-sciences, energy, transport and beyond. There might even be a user-need for a central site — <a href="http://developer.service.gov.uk/">developer.service.gov.uk</a> perhaps? Platform business models are the dominant commercial form of the internet age and the 2011 predictions of how this was expected to develop in the<a href="https://a16z.com/why-software-is-eating-the-world/"> commercial sector</a> have (largely) been borne out, and the infamous ‘<a href="https://nordicapis.com/the-bezos-api-mandate-amazons-manifesto-for-externalization/">Bezos Mandate</a>’ is cited as an example of the type of internal business model transformation that is required to enable that change. Yet the parallel vision of this type of transformation in public services ‘Government as a Platform’ set out by<a href="https://direct.mit.edu/itgg/article/6/1/13/9649/Government-as-a-Platform"> Tim O’Reilly</a> around the same time still feels like a futuristic scenario, albeit a heavily west coast US version of what this might and should look like in the UK.</p><p>To drive both digital public service reform <em>and</em> productivity in the wider tech economy DSIT could embrace this vision that links internal and external digital reform, for example creating a new imperative in the<a href="https://www.gov.uk/service-manual"> Service Manual</a> and<a href="https://www.gov.uk/guidance/the-technology-code-of-practice"> Tech Code of Practice</a> that sets a presumption that government data for different sectors should be made accessible to external entrepreneurs via APIs and developer hubs, with any particular useful datasets containing personal data being exposed through synthetic data equivalents. The DSIT team could champion those existing developer hubs, and help set up a small number of others focussed on improving digital productivity in other priority areas of the economy.</p><p><strong><em>b) Technical competence in policy teams that face into sectors of the economy with digital productivity potential</em></strong></p><p>Another benefit of bringing internal and external digital issues together in DSIT is the opportunity to expand the technical capability of government policy teams that are tasked with driving digital economic growth.</p><p>Many years ago when the digital economy brief sat with Department for (Digital), Culture, Media and Sport I remember a regular cross-Whitehall senior meeting to coordinate interactions with the big platform tech companies. This was treated as a near-diplomatic exercise, and the leading civil servants were steeped in diplomatic skills. But as a result the ‘asks’ made of these companies were debated internally and then communicated with those companies in ways that I fear did not always make deep technical sense. And as a result, I’m not sure these interactions were always met with the degree of respect that government expected. A very noticeable contrast can be made with the way in which a number of deeply technical people have been hired into the government’s AI Safety Institute, and as a result while the relationships might not always be smooth with the frontier model providers, there is a deep degree of respect between the two and there is good reason to believe that the ‘asks’ from the government side at least make technical sense.</p><p>There is a related potential for DSIT to provide a next generation of AI and other technical expertise to the various regulators across Whitehall, many of whom are starting to struggle with the rapid proliferation of AI in the sectors they regulate, which is in turn is starting to radically reshape some of the deeper business models and modes of operation in their sectors. This isn’t the business of AISI, which should be supported to focus on the important frontier AI safety issues, but right now there is insufficient technical leadership in Whitehall about how to regulate in this changing world.</p><p>DSIT could also become an exemplar to the rest of Whitehall about what a new sort of mixed-skill policy team should look like — one where there are multidisciplinary teams with technical people routinely sitting alongside (above, even!) the default policy generalists. When I started my previous government role I observed policy teams dealing with housing and planning were largely unaware of radical changes underway in the PropTech sector, and I expect there will be many more traditional civil service teams curating a section of the economy but only paying attention to the players that have traditionally dominated that sector, rather than those new digital companies that are likely to provide the disruptive productivity boost that the economy desperately needs. DSIT could play a role in reshaping a number of these teams across Whitehall and bringing them together as a network of reforming policy teams — ones where more technical knowledge and a more entrepreneurial approach were actively celebrated, and where a more porous approach could bring in expertise from outside the civil service. The much-mourned previous Cabinet Secretary<a href="https://openpolicy.blog.gov.uk/author/sir-jeremy-heywood/"> Sir Jeremy Heywood</a> was concerned that Whitehall’s policy profession was becoming far too insular and set up a set of<a href="https://openpolicy.blog.gov.uk/what-is-open-policy-making/"> open policy reforms</a> to address this in 2016. There has been progress since then, but let us hope the next Cabinet Secretary returns to this focus on modernising the skills and mindset of Policy Profession and wider civil service.</p><p>One final area where DSIT could make a change is by shaking up the commercial function across government, so it better recognises the economic growth potential from hiring up and coming UK tech companies. The issues for smaller, more nimble and expert companies with the current government commercial approach are <a href="https://www.dxw.com/2024/02/scale-without-impact-government-procurement-today/">well recorded</a>, but these issues are also found across our wider public sector sphere — in local government, policing, the NHS and beyond. Some of us tend to worry about this from the perspective of effective use of public money for public services, but it is also an issue for the British economy at a more strategic level. At <a href="https://faculty.ai/">Faculty</a> we are proud to be part of an ecosystem of exceptional and dynamic home-grown tech companies, we just sometimes wish the state recognised that more in how it buys goods and services and backed companies like us more as we expand and grow.</p><h3>Eight practical things that the new DSIT should do:</h3><ol><li>Align digital to the government processes behind the new missions, and champion this into the Spending Review</li><li>Create an updated narrative for the second decade of digital reform in government that is in tune with the AI age, and use of soft and hard levers to create the exemplars that demonstrate what this looks like in practice</li><li>Activate an army of reforming digital public servants to ensure that new digital service vision gets turned into reality, even in areas that extend beyond direct Ministerial reach</li><li>Create developer hubs and an accompanying ‘Bezos Mandate’ to provide the data on which external digital companies can power productivity growth in key sectors of the economy</li><li>Ensure every digital-facing policy team in DSIT has a strong mix of technical and policy skills, including at senior levels</li><li>Provide technical advice to regulators across government who are starting to struggle with the impact of rapid growth in AI use and shifts in business models</li><li>Create a network of policy teams across Whitehall that face into the sectors of the economy with the strongest potential for disruptive digital productivity growth, and proactively staff them with more technical and entrepreneurial civil servants and external experts</li><li>Shake up public procurement to enable up-and-coming British tech companies to compete and thrive</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f6813b8e8043" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[(Yet more) on a ‘Local GDS’]]></title>
            <link>https://medium.com/@maltbyps/yet-more-on-a-local-gds-b206f3ad1da1?source=rss-d3a206a02212------2</link>
            <guid isPermaLink="false">https://medium.com/p/b206f3ad1da1</guid>
            <category><![CDATA[local-government]]></category>
            <category><![CDATA[local-digital]]></category>
            <category><![CDATA[gds]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Paul Maltby]]></dc:creator>
            <pubDate>Thu, 11 Apr 2024 17:56:07 GMT</pubDate>
            <atom:updated>2024-04-11T18:00:18.896Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/525/1*SF1tlUbkCXkM4A_6-H7X7Q.jpeg" /><figcaption>LocalGovCamp 2017 at Bristol Town Hall &amp; the (soft) launch of the local digital work in DLUHC with the cautionary Reagan quote</figcaption></figure><p>There is a growing and timely public conversation about the potential for a ‘Local GDS’ to help local authorities tackle their digital service challenges and deliver much needed savings, process improvements and better quality outcomes to their residents. Phil Rumens is <a href="https://github.com/LocalGovDigital/localgovdigital.github.io/wiki/Local-GDS">keeping a list</a> (and has also <a href="https://philrumens.blogspot.com/2024/03/local-gds-march-2024.html">added to the discussion</a>), and it includes excellent contributions from <a href="https://medium.com/@jamestplunkett/helping-local-digital-work-b628178df647">James Plunkett</a>, <a href="https://medium.com/@markthompson2442/even-more-thoughts-on-a-new-local-government-digital-service-103bb95c000c">Mark Thompson</a>, <a href="https://peoplefirsttech.org/recommendation-4-put-devolved-and-local-governments-first/">Rachel Coldicutt/Sarah Gold/Dr Natalie Byrom</a>, <a href="https://chiefdigitalofficer4london.medium.com/further-thoughts-on-a-new-local-government-digital-service-24fce01be865">Theo Blackwell</a>, and <a href="https://www.linkedin.com/pulse/do-we-need-new-focus-local-digital-services-jos-creese-ts6ze/">Jos Creese</a> among others.</p><p>I’ve been tempted to comment because when I was Chief Digital Officer at DLUHC <a href="https://x.com/BforBen/status/908983743897112576">I established and ran the local digital team</a> and programme, although I’ve since left government and I am no longer involved. I’ve also worked for a local authority and I’m acutely aware of how different things look from each end of this telescope. In the absence of any senior appetite for anything like a ‘local GDS’ our team in DLUHC attempted to do the best in collaboration with the determined (and growing) group of digital service reformers we knew were already making progress in the sector — often doing so against the odds — and with other supportive colleagues in Whitehall, including GDS.</p><p>It can be quite problematic discussing digital services in an organisation as if they were separate from the day to day service teams that operate the system and the wider government policy and regulatory environment. But it can be useful too, especially as I think its uncontroversial to note that in general councils in the UK have struggled to keep pace even with the progress in digital, technology and data of the wider economy, and also of (parts of) central government. There’s a wider story about the future of councils, including their services, funding and their wider relationships with their local areas and central government. But this isn’t that note, please forgive me.</p><p><strong><em>Recent experience</em></strong></p><p>Its easy to start with the negatives, but I hope any discussion about a ‘Local GDS’ starts with a proper recognition that the sector <em>has</em> made progress in digital services, especially for new and niche services. It has a much stronger digital workforce and leadership across the sector than it had a decade ago, and it is improving all the time. There are happily far too many places to list by way of example, but there is great digital and data work stretching across very different local authorities including Hackney, Camden, Newcastle, Adur and Worthing, Norfolk, and at a regional level via Loti and the Greater Manchester Combined Authority. (Making lists like this is so unfair — there are many, many more that could and should be listed in any local authority digital hall of fame).</p><p>The approach when we started the DLUHC local digital team was to a) bring the sector together around a sensible, modern vision for reform via the <a href="https://www.localdigital.gov.uk/declaration/">Local Digital Declaration</a>; b) find and back the reformers (like those in <a href="https://localgov.digital/">LocalGovDigital</a>) in the sector, with plenty of meet-ups in person and online to share ideas and give strength in realising we weren’t battling alone; and c) establish a programme of work that concentrated on doing what ‘only’ central government could do to help move things forward.</p><p>The focus in the early days of that central activity was to use a <a href="https://www.localdigital.gov.uk/fund/">Local Digital Fund</a> to support groups of local authorities to develop re-usable services in more new/niche categories of activity, rather than take on the behemoths of the big domain services like social care. As things developed we leaned in to support better cyber protection in the sector (to prevent more ransomeware attacks as seen in <a href="https://www.bbc.co.uk/news/uk-england-tees-53662187">Redcar</a> and <a href="https://news.hackney.gov.uk/cyberattack/">Hackney</a>), invested in digital training for officers, and we picked off one of the big set piece domains where DLUHC itself had Whitehall policy responsibility — <a href="https://dluhcdigital.blog.gov.uk/category/digital-planning/">planning</a>.</p><p>To my ears this experience of the last few years in both local authorities and Whitehall feels too absent from some of the recent discourse about a new ‘Local GDS’, but there is much to learn from it, not least as the programme is undergoing a well-scoped formal <a href="https://www.gov.uk/government/publications/local-digital-evaluation-scoping-study">evaluation</a>. There is hard-won expertise in the DLUHC local digital team about how to curate technology, relationships and (critically) the trust that will be vital if we are collectively to have more impact. There’s a long way still to go of course, but there valuable momentum has been built and any new ‘Local GDS’ would be wise to work in a way that harnessed this knowledge and energy at both local and central levels.</p><p><strong><em>The big challenges and opportunities</em></strong></p><p>From my experience there are two big challenges that any reform programme needs to seriously grapple with:</p><ul><li>First, the main domain services and software in local government (social care, revs/benefits, waste, highways, housing etc) have been largely impervious to reform, and there is an imperative to act given this is where most of council budgets go. Now is the time to invite competition into these big domains and bring an end the closed, full-stack oligopolies/monopolies that are endemic in the sector and which frustrate the current and future delivery of efficient services. This was done at scale in central government after 2012 as a result of the <a href="https://www.gov.uk/government/publications/directgov-2010-and-beyond-revolution-not-evolution-a-report-by-martha-lane-fox">GDS revolution</a> and the benefits are considerable.</li><li>The second big challenge is that, as with other parts of the public sector, there remains a deep lack of digital/data understanding in leadership teams, both among officers and councillors. Any progress in ‘digital’ can only come about as part of a broader transformation of services, processes and culture in the wider organisation — especially given that in reality technology choices in the big domain services are still driven by Service Directors, not the CDIO/CTO.</li></ul><p>On the upside, there are some great examples to draw from now in how to reform those big monopoly service domains, although these are not known by many outside the sector: <a href="https://dluhcdigital.blog.gov.uk/category/digital-planning/">local digital planning</a> and <a href="https://localgovdrupal.org/">local drupal</a> (powering council websites) are two good examples. In both of these examples there was genuine and healthy collaboration between council reformers on the ground and the DLUHC digital team in central government who played a role as funder, backer and guide through the resulting market disruption.</p><p>The next stage of reform in local government needs to continue this mission but broaden the scope beyond planning and websites. It needs to do the hard work to tackle those other main domains, especially those where there is significant potential to unlock resource and reform at a system level — not least in adult and children’s social care.</p><p><strong><em>What would a next phase of reform look like?</em></strong></p><p>To do this requires a practical partnership formed of (multiple) councils, the Whitehall departments responsible for overseeing the domains in question (plus the associated regulators), and a central digital driving force — whether we call it ‘Local GDS’ or not. Central support currently comes via the local digital team at DLUHC, but if it meant extra political support and funding this leadership could come from elsewhere, so long as those running this entity fundamentally understood the nature of local government variation and sovereignty and wanted to work in a collaborative way with the sector.</p><p>The specific plan for the type of reform might need to flex in each domain, and that’s probably fine. For me, I think the critical elements would need to be:</p><ul><li>A collaboratively written vision along the lines of the <a href="https://www.localdigital.gov.uk/declaration/">Local Digital Declaration</a>, driven by user-need and efficiency across each whole end-to-end service, solving user problems first time, making the most of data and AI to improve service optimisation and efficiency.</li><li>Working on the assumption that while a handful of councils may genuinely need to configure an individual services in a genuinely novel way most councils will be pleased to (re)use good services that are available in a few patterns of variation (e.g district/unitary/county; rural/urban). Local drupal offers an intersting glimpse of the possibilities here.</li><li>That the services are built in an interoperable, modular way using modern platform technologies, including the use of AI to drive operational efficiencies. As <a href="https://medium.com/@markthompson2442/even-more-thoughts-on-a-new-local-government-digital-service-103bb95c000c">Mark Thompson says in his post</a> there’s no reason to bespoke much of these services each time in each area, whether by internal digital teams or external providers. The ‘Local GDS’ central team would ensure that there was sufficient componentisation as each digital service for these main domains develops, which should be taken from an existing thriving digital market, especially lower in the stack. We shouldn’t see any internal public service digital teams making components that already, or could, exist commercially — let’s all remember the power of <a href="https://learnwardleymapping.com/">Wardley Mapping</a> to work through where we should be using commodities and products and what should be custom built.</li><li>This central ‘Local GDS’ team should be thinking commercially about both individual components and the service as a whole. Components need to be thought about as being extended and re-used across different services within an organisation: so e.g. security features or casework modules should be designed for one service (e.g. adult social care) but then be expected to be re-used in other domains (e.g. charitable services grants funding) that have similar basic functions. But it is a local matter for CDIOs/CTOs to organise their architecture and platform strategies — a remote central team will be a disaster if it tries to plan all this out in each council from an ivory tower. However, a central team should be working hard to both ensure there are the standards to enforce a fundamental tech interoperability and also the commercial pricing and marketplace for new competitors for those individual components — at least during an early phase of reform where stronger intervention is necessary to address the current market failures.</li><li>But the big prize for this central ‘Local GDS’ team and its stakeholders is to craft a more functional digital service market at the <em>whole service level</em> — and there is a real risk we replace one set of near-monopoly providers with a second generation ‘modern’ monopolies that are on a fast-track to becoming legacy. It is tricky to work through whether and how competition and flexibility works at both the component and whole-service model, and how to organise this within both councils and at a whole-system national level. My experience in DLUHC surfaced sometimes conflicting incentives and trade-offs between an approach at a component level and an approach at a whole-service level, and there was little useful wisdom on how to craft these imperfect digital markets within the existing commercial functions in government.</li></ul><p><strong><em>A plan to tackle the big service domains in local government</em></strong></p><p>At a more detailed level how might such a ‘local GDS’ function work in relation to these big domain services? There are clearly many variations possible, but to get the collective conversation into more of the particulars here is a hopefully useful straw model:</p><ul><li><em>Phase 1:</em> For each of the significant domain services, user-centred design processes would be used in the normal way to create a ‘competitor service’, drawing on mostly pre-existing commercial components, with the agile development funded by central government and a small group (e.g. x6) early adopter councils. The resultant open source service would be organised as a collaborative legal entity, owned by councils, central gov (and maybe civil society organisations where appropriate — e.g. <a href="https://opendigital.coop/">LocalDrupal is owned by a cooperative</a>). The services would be built with the assistance of in house teams and digital service agencies, who would hired with an expectation they would play a role in phase 2</li><li><em>Phase 2:</em> Central government would then bring together a wider group of local authorities to procure a ‘live’ version of the service from a commercial provider, presumably with at least one part of the provider coming from the agency that helped build the service. This ‘live’ version would consist of a service package around the fundemantal open source product which would be offered to councils, incorporating bug fixes, security patches and other incremental improvements back to the open source service, plus other elements like first-line support and training (a model akin to <a href="https://wiki.openstreetmap.org/wiki/Mapbox">Mapbox vs OpenStreetMap</a>).</li><li><em>Phase 3:</em> After a period of time, set out in advance (e.g. three years), central government would help orchestrate a second round of commercial bidding, whereby at least two and possible more large groups of councils would come together to buy the commercial service package from <em>different</em> providers. The original open source commercial entity (comprising central and original local government members) would decide what elements of the service would need to remain as common (open source) standards, potentially mandated through legislation or guidance as with <a href="https://www.localdigital.gov.uk/digital-planning/policy-and-legislation/">Digital Planning</a>, this would include e.g. the API standards for accessing different parts of the stack to retain a competitive market and public service utility, the data standards (including open data publishing), perhaps the regulatory requirements for interfacing with the service’s government regulators. At this point the old-world previous incumbent provider must be required to meet these standards if it is to remain available in the local government marketplace. The expectation is that at this point the new service companies would each fork the original open source model and make their version proprietary (although they could choose to keep their version open source if they wanted). This would be designed to create the commercial incentives to invest commercially in the development and iteration of the original service, to keep a momentum for ongoing reform and end or reduce the amount of additional public subsidy.</li></ul><p>These reforms should tackle head-on that there needs to be genuine jeopardy for incumbent (near) monopolists: rival services will be built and technical standards will be enforced with the accompanying threat of legislative/regulated entry to a market, as alreadydescribed in <a href="https://www.legislation.gov.uk/ukpga/2023/55/enacted">clause 87 of the Levelling Up and Regeneration Act 2023</a></p><p>In addition to creating genuine alternatives in the market there are some other cucial blockers/enablers that need to be tackled collectively/centrally and which need to be in a remit of any ‘local GDS’, including updating digital procurement practice, API and data standards, cyber protocols and assessments, and attracting digital talent. The good quality recent work of the DLUHC local digital team on <a href="https://dluhcdigital.blog.gov.uk/2024/03/14/developing-a-service-to-assess-and-understand-cyber-resilience-in-local-government/">cyber</a> and <a href="https://dluhcdigital.blog.gov.uk/2024/04/10/scaling-data-standards-to-support-local-government/">local standards</a> are useful reference points for what this could look like.</p><p><strong><em>Things to avoid</em></strong></p><p>There are, however, also some things to be avoided:</p><ul><li>We should all be on our guard against creating a daft centralising force that assumes that councils are rubbish and that they all just do the same thing without any fundamental local difference — ‘just do it once for them from the centre, and they can add their logos if they want’. <a href="https://www.linkedin.com/pulse/do-we-need-new-focus-local-digital-services-jos-creese-ts6ze/">Jos Creese</a> flags this in his post, and I’ve seen folks in central government misjudge this issue, horribly, time and time again.</li><li>We all love open source, it should go without saying. But I’ve seen too much commercial naivity in the local digital sector about the business models behind local government software, and the large amount of time and ongoing investment required to keep software safe and up to date. There are nasty market failures in the local government software sector, and some of these are likely structural in nature. But there are plenty of commercial organisations that do and would add serious value to the sector and we collectively need to do better than wishing away market complexity with a type of hippy open source magic wand.</li><li>The public sector need to avoid having endless meek discussions with embedded (near) monopoly software providers about opening up their services and predicating the whole mission for local digital reform on an (ignored) set of data standards — the NHS has wasted years doing this and remains riddled with terrible software. Monopolists are going to monopolise and extract rents, its how it works. Yes of course the public sector needs engage all suppliers at all stages, and it should reach for the the incentives and routes for existing providers to pivot and the make the most of their domain and technical talents in a new world. But commercial naivity in this space will not be rewarded.</li></ul><p><strong><em>Eyes on the prize</em></strong></p><p>There are big opportunities if digital/AI reform in councils is done well, and there is a helpful momentum towards doing better in this area than we have done collectively in recent years. Local authorities are a £2bn+ technology market afterall, and play a critical role in the delivery chain of many national-level public service priorities — not least the NHS.</p><p>Yet there is a significant risk any ‘local GDS’ or equivalent will get horribly distracted. There are thousands of individual services in operation across local government, hundreds of individual councils, and strong and necessarily divergent political agendas across the sector. People who work in councils have views about their future and are often battling against unhelpful structures and processes. Councils are not simply transactional delivery agencies, they are both curators of and participants in their local community, democracy and economy. There will be fancy new technologies that come and go (blockchain! virtual reality!) that bedazzle but hold out very little opportunity for scaled efficiency reform. All of these and more will provide a mix of distractions, worthy ideas, and side-quests that will be difficult to refuse.</p><p>But the overriding issue is how to deliver mainstream operatinal council services at the highest quality and lowest cost possible, using the digital/data approaches that all other sectors, globally, are already employing to drive similar change. In councils this means having a lazer-like focus on that handful of big end-to-end services that drive much of their activity and costs, and on the software, processes and use of data that sits behind them.</p><p>We might create a ‘Local GDS’, and personally I’m rather keen for something with more clout (and money) in this space regardless of what it is called and where it sits. But the bigger issue is what should (and shouldn’t) those interested in local digital service reform focus on. My answer is creating user-led, end-to-end, geunuinely digital/AI ‘challenger’ rivals to the embedded software and its acommpanying service processes in those main domain areas. Without this, we’ll all be collectively playing on the margins and those vital local services and the people that depend on them will continue to struggle.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b206f3ad1da1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Quick thoughts on GDS’ ‘What next for Gov as a Platform’ post]]></title>
            <link>https://medium.com/@maltbyps/quick-thoughts-on-gds-what-next-for-gov-as-a-platform-post-007c01d97162?source=rss-d3a206a02212------2</link>
            <guid isPermaLink="false">https://medium.com/p/007c01d97162</guid>
            <category><![CDATA[gds]]></category>
            <category><![CDATA[digital-government]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Paul Maltby]]></dc:creator>
            <pubDate>Tue, 02 Jan 2024 18:33:17 GMT</pubDate>
            <atom:updated>2024-01-05T10:32:30.648Z</atom:updated>
            <content:encoded><![CDATA[<p>I’m really pleased to see <a href="https://www.martinlugton.com/what-next-for-digital-government-and-government-as-a-platform">this blog post</a> from <a href="https://www.linkedin.com/in/martin-lugton/">Martin Lugton</a> about ‘what next for digital government and government as a platform’. It is never as easy as non-government people think to share early thinking as a civil servant, even if in my experience when I was able to do so I nearly always found it helped improve the quality of the resulting work. So thanks to Martin and the widerleadership team for enabling him to put this out there.</p><p>From a personal perspective there’s much to admire and agree on in Martin’s post, but I have an overriding worry that the paradigm of digital government in the UK has got rather stuck in the (super-admirable) 2012 GDS reforms and I think that model is running out of road, potentially crowding out other important elements. Here’s three quick thoughts about what I might add, drawing on my time in government, latterly as CDO in DLUHC, and also from a year spent at <a href="https://faculty.ai/">Faculty</a> observing and working with a much wider range of government teams:</p><ul><li>First, we should make it easier to identify and deploy mixed (state/commercial) components that make up any modern digital service, remembering few (any?) of these components really need to be built by the government itself these days. Despite the GDS/CDDO split, which probably makes it harder than it should be, a next GDS product and platform strategy could better acknowledge this mixed commercial reality and and make it easier for service teams in departments to find, procure and implement these components in ways that best meet user (and taxpayer) needs. <a href="https://youtu.be/ZzPU6Pdw05s?feature=shared">This Government as as Platform video </a>from almost a decade ago got the point about the wide range of platform products, even if it too rather implied GDS should ultimately build them all. We should ask why are digital service teams in government still scrambling from scratch each time we build a service to understand what components are available in the market to do particular jobs, and how best to connect (and procure!) them? How many service teams in government still cede these decisions to their digital agency partners without having much of a view or a say in the components that are chosen? And how are we helping the government digital services market evolve to understand and develop healthy, competitive horizontal markets for these components across our individual vertical service lines?</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/866/1*B7F8rfSOE7PUKGPOjdaE9w.png" /><figcaption>Screenshot of early GDS ‘Government as a Platform’ video, via YouTube</figcaption></figure><ul><li>Second, where are the common service patterns that lie above the component tier, and why is it so hard to find these as a digital service team in government and to reuse these or rapidly/easily procure their individual elements? Why do we have so many different systems bespoked for essentially the same type of administrative function, including within each department? Sure, government is complex, but if you squint there are only a small number of basic <em>types of services,</em> but these are not mapped out centrally and their common building blocks (including their user research, service design &amp; architecture patterns and technology components) are hard to identify and reuse by other government digital teams. I agree that there may be a few examples where we really do need GDS to intervene and build a whole thing, and personally I’m inclined to agree with Martin that GOV.UK forms with more automated service wrappers for the long tail of low-volume (currently) PDF forms across the state is one of these. But generally I think we should be super wary of assuming a GDS hands-on ‘build’ role, rather departments should have access to a centrally curated, technically sophisticated set of patterns and connectors and procurement routes that they want to make use of because it is quicker and easier, like they already do with the <a href="https://design-system.service.gov.uk/">GDS design system</a>.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3_umuC5V4kjlcx73QgLwfA.png" /><figcaption>A rough illustration of possible patterns of types of government services above the component layer, highlighting that components and infrastructure will often stretch across a single ‘vertical’ service</figcaption></figure><ul><li>Third, where are the ‘missing’ elements of services that have already been reformed in the 2012 GDS model, that the centre of government should be using its cutting edge expertise to lead on and help others in departments understand and implement within their services? For example there is still precious little scaled adoption of AI (including ‘traditional’ AI and not just GenAI) within transactional and caseworking services which could do much to help reduce the blight of backlogs, and potentially to do so quickly. There are also few examples of more sophisticated digital interaction with citizens as part of services, of the type we can see being developed in <a href="https://media.localdigital.gov.uk/uploads/2023/11/28105757/PropTech-Innovation-Fund-Combined-case-studies.pdf">DLUHC’s digital planning work</a>, and we see in other countries via tools like <a href="https://pol.is/home">pol.is</a> and which a number of us recently saw demo’d in <a href="https://www.itworldcanada.com/article/diia-ukraine-implements-world-leading-digital-government-in-the-middle-of-a-war-whats-our-excuse/502597">‘Diia’</a> the Ukrainian government app (used for deadly serious needs, in addition to deciding Eurovision entries). And where are the new business models unlocked by new technologies such as GenAI that should be reshaping the strategic <em>policy</em> agenda of government, and not just be hidden in the technology basement downstream from the big decisions? A next generation GDS/CDDO could be highlighting these, helping departments to ship practical examples, and writing the updates to the service manual to enable more rapid, scaled deployment of these next generation elements in the way that the early GDS did with user-centred product thinking.</li></ul><p>There’s no doubt so much more to say here, for instance about the importance of continuing to improve the (where appriopriate) external accessibility of data through APIs which is crucial not only for the original <a href="https://www.oreilly.com/library/view/open-government/9781449381936/ch02.html">O’Reilly-esque</a> articulation of government as a platform (rather than <a href="https://www.computerweekly.com/opinion/Government-as-a-platform-or-a-platform-for-government-Which-are-we-getting">platforms for government</a>). But which is also as something needed if we are to encourage an AI productivity boost across the wider economy, again something I think that the digital planning work in DLUHC is getting right. But I’ll leave these thoughts for a later date.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=007c01d97162" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Government as a Platform 2.0: How government publishing needs to change in a world of generative AI]]></title>
            <link>https://medium.com/@maltbyps/government-as-a-platform-2-0-how-government-publishing-needs-to-change-in-a-world-of-generative-ai-82eecd9d9222?source=rss-d3a206a02212------2</link>
            <guid isPermaLink="false">https://medium.com/p/82eecd9d9222</guid>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[digital-government]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[large-language-models]]></category>
            <category><![CDATA[govuk]]></category>
            <dc:creator><![CDATA[Paul Maltby]]></dc:creator>
            <pubDate>Fri, 09 Jun 2023 10:49:58 GMT</pubDate>
            <atom:updated>2023-06-12T10:01:35.548Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/854/1*je2uJB7KD3ms3GQhtH--nw.jpeg" /><figcaption><a href="https://flic.kr/p/d2Kmbs">https://flic.kr/p/d2Kmbs</a></figcaption></figure><p>There are many potential use cases of large language models (LLMs) and other types of generative AI in public services, but few will be as visible to the public as a potential new generation of chatbots used to interact with government information or services. In the short term, these will no doubt be an adjunct to the typical way of accessing GOV.UK directly in the browser or through a search engine.</p><p>But LLMs may soon dominate the way people interact with government information if they offer people a route to find what they want more efficiently and enable them to ask clarifying follow up questions.</p><p>On a personal level, I was recently surprised how effectively ChatGPT (using the less powerful GPT-3.5) gave me effective structured answers about a niche area of tax policy. The ability to ask it follow-up questions to probe particular areas was a revelatory new experience of interacting with government information.</p><p>Many others are already doing similar things. Even though, as <a href="https://faculty.ai/openai/">our partners at OpenAI</a> are keen to remind us, these models are <a href="https://www.theguardian.com/technology/2023/mar/17/openai-sam-altman-artificial-intelligence-warning-gpt4">reasoning engines rather than fact databases at heart</a>. Out of the box, the information they know is already a couple of years out of date and prone to hallucination, which emphasises the need to incorporate validation and checking tools around the models.</p><p>In the near future we might expect that such chat interface services, designed specifically to reduce the risks around out of date material and hallucination, will not only be offered by government directly. People may expect information to come to them via their own LLM-based ‘AI assistants’ (or whatever we will come to call them), likely embedded in their browsers or their day-to-day office technology.</p><p>It is also possible that information will need to flow from government through an intermediary LLM service via a company or charity in the course of answering a wider and more bespoke question for a citizen.</p><p>We are starting to understand that that this change will require a rethink about how core information is provided by government. And it might lead to a new imperative for public services to do better in how they expose information, rules and data.</p><h4>GaaP in a time of LLMs</h4><p>A first generation ‘<a href="https://watermark.silverchair.com/inov_a_00056.pdf?token=AQECAHi208BE49Ooan9kkhW_Ercy7Dm3ZL_9Cf3qfKAc485ysgAAApMwggKPBgkqhkiG9w0BBwagggKAMIICfAIBADCCAnUGCSqGSIb3DQEHATAeBglghkgBZQMEAS4wEQQMlX2H_kH_hGX3cCG9AgEQgIICRlS24mJUa9FRJmdF7e0q5QnPp-AiaOf6v-edZzNqDNGWBxbJptYMObbdX9Q_0w7wruvMScHLh0JmfaS4hFoDEmB2gAZWyVDHQfenN_gl4MvW3EPChvN1FSi_mrvDgiAWlUys0i4wJT5Qb3HKo3Uk1IWkjVYWIQeeoO1cGv687AmyLPWDGUed6O2Zfyg8Ns6CE_mhL4Nl_r1hMuys1iTgpyFK5fw-wLbd48XeTJ8EAe3vcadhtnKn5uPe0PBV6Yb_RCIndKtMKqa1fs0FtanQzotWn3aWnNZBcsXGuU3nsd9iV47UEATj8iekVZaDRuY1w1em-BHrJF3Iz_1m1skGDwbHZPTbOYRk8_iYgx19faVLFx35t4gOcDVuUEGZ8TtEWx9ftgkzjC97oYCNVUEjNsXv4gy-XRFq8XSoLCH6wajcDAgAyPfx7gSH-jqvO-z0eaBuK0MYP3oBkw2p4_skswj3pFiK4OWM0vlaRZiV4uaHA3ZcTw-YRfrX_9RIvDTyarfnlTXoE4vy23XNVHWI4zIVObDNgdRhZf6E0Lobe9tOPnzpYNfiSapPkoswvzz9o_DerPQLx4Tp1rRPMK9IR6-_iRRvRJuSktL1LenjHc5iV4x4iiQl1btZVmXSQ4TMlKzGrU1Xu4yn0xrDhAvM-tmkVHo1bz7ouADTUbxjybI4WHFAX6Tqz1ZWOxAaiKwySJXTKNnlsp_tTKcJTDP49o13VGzLTYgEZhsgYhzzVttPV_chXTKye_eHRnRLaaQYl06NaoCa7A">Government As A Platform</a>’ outlined by <a href="https://twitter.com/timoreilly">Tim O’Reilly</a> in 2010 was designed to expose data so that others could use it to build digital services on. It could have been a boon for public services and economic growth. But it largely failed to penetrate senior government consciousness and never really caught on in the UK beyond some interesting use cases such as HMRC’s <a href="https://www.tax.service.gov.uk/making-tax-digital-software">Making Tax Digital</a>, DVLA’s <a href="https://developer-portal.driver-vehicle-licensing.api.gov.uk/availableapis.html#current-available-apis">driving licence developer portal </a>and <a href="https://www.planning.data.gov.uk/map/?">DLUHC’s digital planning reforms</a>.</p><p>By contrast, in a time of LLMs, the need to expose information to be read by these next generation digital services, rather than people direct, could be much more existential for the government. This idea from 2010 may require a new urgent twist — we might call it <strong><em>Government as a Platform 2.0</em></strong><em>.</em></p><p>Alongside many others in government, we are keen to get beyond the current wild west of in-office ChatGPT use, rather abstract policy debates, and get on to more learning-by-doing through more considered applications. LLMs are powerful and highly deployable, but in the end, they are just digital services.</p><p>There is still much to be learned about their real world annoyances, errors and ability to meet user needs. The tried and tested way to do this is by rolling up our sleeves and getting on with practical work.</p><h4>Mixing fine-tuning and in-context learning</h4><p>Even in these early days of next-generation foundation models we are learning quickly about how to get the best out of them — through more sophisticated prompt engineering and by squeezing the most out of the limitations (and costs!) of information passed to the models in fine-turning and in-prompt contextual learning.</p><p>At <a href="https://faculty.ai/">Faculty</a>, we have seen good results from using natural language processing algorithms to understand what type of content is contained in which sections of long documents, so that we only pass the right sections into the GPT-4 model. Passing a thousand page document into the GPT-4 prompt costs around £15 a time currently, which can soon add up for a web-scale service.</p><p>As early prototypes get underway we should actively monitor how to get the best out of them. It is widely understood that we have different options for getting specialist information out of the models. From using them out of the box as is; to fine-tuning them so they have more specific background information and stylistic understanding of a particular type of knowledge; and through providing additional specific contextual information in the prompt.</p><p>The latest models are seriously increasing the capacity of the latter to 24,000 words for OpenAI’s GPT-4 and 75,000 words for Anthropic’s Claude. Our working assumption is that use cases will often require a mix of fine-tuning and in-context learning through the prompt. Although the playbook has yet to be written for when particular combinations prove more effective.</p><h4>Preparing source material for LLMs</h4><p>The less-commented on variation is what needs to be done, if anything, to the specific source material to get the best out of it via an LLM.</p><p>As a digital profession, we have come a long way in improving how we enable people to get the most from GOV.UK, and understand the benefits of <a href="https://www.gov.uk/guidance/content-design/what-is-content-design">content design</a> to improve how to make information clear and useful for people. We have reusable patterns such as <a href="https://design-system.service.gov.uk/patterns/step-by-step-navigation/">step-by-step</a> navigation through more complex transactional services, part of a body of knowledge that is the commendable <a href="https://design-system.service.gov.uk/">GOV.UK design system</a>.</p><p>But we need to start to create a similar body of knowledge for how LLMs best access information via the prompt. For instance, it is plausible that articulating those step-by-step processes will also improve the ability of an LLM to give clear information back to a user compared to the (still too common) wordy civil-service-ese that many government web pages and policy documents are still written in.</p><p>But it is not a given — maybe LLMs do <em>not</em> need good content design to present clear information back to people. Indeed, our early work with the latest generation of LLMs suggest an incredible ability to create clear, sophisticated and accurate summaries from messy, lengthy and even technical information.</p><p>As a result, it feels like a strong early use-case for LLMs in government is for those parts of government such as regulators and operational arms-length bodies to better communicate to both internal and external users the rules for specific use cases based on their own complex operational and technical guidance, legislation and even the newly accessible (but — frustratingly — not legally useable by algorithms) <a href="https://caselaw.nationalarchives.gov.uk/">case law </a>from The National Archives.</p><p>We need to also consider how best to prepare materials for fine-tuning the models, not just for passing information via the prompt. Do we need GOV.UK instruction tuning datasets which hold anonymised data on all the tasks that people try to achieve online? Do we need specific datasets to fine-tune models on specific domains (like Ministerial Correspondence, Case Law, technical regulations)? And do we need to start creating datasets of <a href="https://huggingface.co/blog/rlhf">human feedback </a>for government-specific material?</p><h4>Disrupting data curation best practice</h4><p>What we really do not know yet is whether, or to what extent, LLMs would benefit from having this information structured more formally in advance. Digital and data teams are accustomed to expecting they will have to put hard work in up-front to curate data to get the most out of it, but it is plausible that LLMs will significantly disrupt what best practice looks like.</p><p>For example, the <a href="https://www.digitalmarketplace.service.gov.uk/digital-outcomes-and-specialists/opportunities/14498">Open Regulation Platform </a>from the Better Regulation Executive is aiming to make it simpler for businesses to know which regulations apply to them and how they can comply with any requirements by processing the UK’s statute book and other regulative data sources into an enriched machine-readable dataset of business regulations that is publicly accessible through an API.</p><p>In a different but parallel approach my previous team in DLUHC are exposing the underlying rules and data to make the planning process less risky and ambiguous to its users in the<a href="https://opendigitalplanning.org/"> Digital Planning programme</a>. It is surprising to many that our planning system is more rules based — ’‘if-this-then-that’ — than many expect, especially for high volume, low-complexity cases that swamp planning offices. But the rub is that those underlying rules and data feel ambiguous because they are not well exposed and as a result are hard to find, trust and use.</p><p>These types of projects, designed to structure and expose data and rules, have been slow to scale due to their upfront complexity and cost. Perhaps they are now <em>vital</em> to the next phase of government publishing, especially for passing material to prompts.</p><p>The counterfactual could also be true — these large language models trained on the messy internet might be better suited to dealing with similarly unstructured data for fine-tuning or for sending as part of the prompt — so long as the rules and data are exposed somehow. Perhaps well-intentioned (expensive) schemes that are trying to better order data to be machine-readable are now over-cooking what is required?</p><p>An additional point to consider is that the performance of the LLM model itself is only one part of the equation when these tools are deployed as more sophisticated digital services.</p><p>Developments such as <a href="https://python.langchain.com/en/latest/index.html">LangChain</a> and equivalents bundle components around the core LLM such as the ability to access the web and validate sources which improve the accuracy and usability of LLMs. Considerations about how government publishing needs to change to optimise LLMs will also need to take into account these emerging validation and performance enhancing tools too.</p><h4>A collective approach to fine-tuning LLMs</h4><p>The answers to all this will only come from getting on with practical work and early prototypes. As these first LLM projects get underway in and around public services we should encourage a collective effort to measure and document the performance, user satisfaction, costs and time of these different approaches to how to best structure the inputs to fine-tune an LLM. And how to best pass the information to the model via prompts.</p><p>In the likely scenario that some up-front work is required to get more usable results for users, then this second-wave <strong><em>Government as a Platform 2.0</em></strong> might be something we hear more of in the months and years to come.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=82eecd9d9222" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Prototyping a One Team Government manifesto]]></title>
            <link>https://medium.com/@maltbyps/prototyping-a-one-team-government-manifesto-1596d99d3f57?source=rss-d3a206a02212------2</link>
            <guid isPermaLink="false">https://medium.com/p/1596d99d3f57</guid>
            <category><![CDATA[innovation]]></category>
            <category><![CDATA[government]]></category>
            <category><![CDATA[digital]]></category>
            <category><![CDATA[policy]]></category>
            <category><![CDATA[oneteamgov]]></category>
            <dc:creator><![CDATA[Paul Maltby]]></dc:creator>
            <pubDate>Wed, 28 Jun 2017 16:26:05 GMT</pubDate>
            <atom:updated>2017-06-28T16:26:05.941Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*avVmMI_C1Xp8-Ly3h3TVgA.png" /></figure><p>There are always going to be multiple versions of more effective working between policy and digital, reflecting the breadth of activity in both disciplines. The space between <a href="https://gds.blog.gov.uk/2016/04/18/what-we-mean-by-service-design/">digital service design</a> and the operational end of policy is perhaps the place where a more seamless joining is possible, and colleagues have already started to envisage how these disciplines <a href="https://sfadigital.blog.gov.uk/2017/03/24/dont-bring-policy-and-delivery-closer-together-make-them-the-same-thing/?utm_content=buffer5ae34&amp;utm_medium=social&amp;utm_source=twitter.com&amp;utm_campaign=buffer">could even merge</a>.</p><p>For others in other policy officials wanting to become more digital the relatively recent work on <a href="https://openpolicy.blog.gov.uk/what-is-open-policy-making/">open policy making</a> is a good place to start, although this was concerned with a broader question of policy innovation rather than just ‘digital’, and sadly the Cabinet Office team was disbanded in early 2016. The team created an <a href="https://www.gov.uk/guidance/open-policy-making-toolkit">open policy making tool kit</a> with practical tips and illustrative examples from across government, and the team led by <a href="https://twitter.com/nybergmaria1">Maria Nyberg</a> created the following useful summary of what was meant by <em>open policy making</em>:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*S1G9xVyxL-v7NOoV." /></figure><p>To some extent the work of the original unit evolved through its sister team<a href="https://openpolicy.blog.gov.uk/"> Policy Lab</a>, a small Cabinet Office unit that makes available design, digital and data techniques to policy makers across government. Policy Lab drew on the experience of <a href="http://mind-lab.dk/en/">Denmark’s MindLab</a> and a <a href="http://www.nesta.org.uk/publications/i-teams-teams-and-funds-making-innovation-happen-governments-around-world">worldwide cluster of innovation teams</a>, and under the leadership of <a href="https://twitter.com/AndreaSiodmok">Andrea Siodmok</a> has contributed <a href="https://www.slideshare.net/Openpolicymaking/policy-lab-slide-share-introduction-final">its own style and methodologies</a>.</p><p>Digital policy making should also draw heavily from general good practice in policy making; the civil service <a href="https://twitter.com/PolicyProfUK">policy profession</a> has been improving training and effectiveness under the lead of its head, <a href="https://twitter.com/PolProfHeadUK">Chris Wormald</a>. And the original <a href="https://www.gov.uk/design-principles">GDS design principles</a> remain deeply relevant for more general policy makers despite being focussed on delivering digital services.</p><p>And in terms of mindset and energy, there is much for policy officials to learn from the organic community around digital public services seen through the meetup culture of <a href="https://www.ukgovcamp.com/">GovCamp</a>, <a href="https://teacamplondon.com/">TeaCamps</a>, <a href="http://odcamp.org.uk/">OpenDataCamp</a>, <a href="http://localgovdigital.info/">LocalGovDigital</a> and similar. It is striking that there is nothing similar to this from the policy community (could #policycamp work?). There are a few examples of policy innovation communities further afield, although with limited success — for example the early years of <a href="https://dinobusters.org/if-theres-something-strange-in-your-government-who-youre-gonna-call-dinobusters-free-interpretation-of-the-ghostbuster-song/">Dinobusters</a> in the Dutch government before their parting of ways, and <a href="http://www.thegovlab.org/">GovLab’s</a> efforts to form an international <a href="http://www.thegovlab.org/project-network-of-innovators.html">network of innovators</a>.</p><p>So if we were to offer a blueprint for what doing or being ‘digital policy’ in government is, aiming for a government of (and not just on) the internet, what might it look like? What would be its outlook, its practices and scope? Below is a first attempt at a digital policy One Team Government manifesto, offered to help prompt a longer and more collaborative discussion. It is written from the perspective of what being a digital policy maker could look like rather than something that aims to accurately describe the full range of roles across both the digital and policy professions, or of a newly merged single role (though it may be that such a broad perspective is possible). I’ve <a href="https://github.com/paulmaltby3/OneteamgovManifesto">posted the text on GitHub</a> to make it easier to iterate, and I’d welcome contributions.</p><blockquote><strong>One Team Government manifesto — prototype</strong></blockquote><blockquote><strong>1. Be curious, practical, reforming</strong></blockquote><blockquote><strong><em>We are reformers. </em></strong>We embrace change and define ourselves against those who are reluctant or fearful of it. We serve Ministers elected by the public and we follow the <a href="https://www.gov.uk/government/publications/civil-service-code/the-civil-service-code">civil service code</a>.</blockquote><blockquote><strong><em>We are curious.</em> </strong>We are familiar with the <a href="https://github.com/paulmaltby3/digitalpolicyreadinglist">best thinking and writing</a> of the internet age, and are hungry for more knowledge and experience.</blockquote><blockquote><strong><em>We roll up our sleeves.</em></strong><em> </em>We mess around with data, code and digital tools ourselves, even though that is not always our day job. This first hand knowledge helps us cut through the hype of new fads and fashions.</blockquote><blockquote><strong><em>We are the future of the civil service. </em></strong>The internet came for retail, finance and the media and it will come for us unless we are prepared. We know it is <a href="https://medium.com/doteveryone/what-a-digital-organisation-looks-like-82426a210ab8">not okay to be ignorant about how our modern world functions</a>.</blockquote><blockquote><strong><em>We understand that digital changes the fundamentals.</em> </strong>We see that digital is reshaping the foundations of society, and understand this affects how we do policy. It is a category error to think of digital as merely a new add-on, something to be considered after core policy work has happened.</blockquote><blockquote><strong>2. Have a deep regard for the citizen experience of a policy or service</strong></blockquote><blockquote><strong><em>We are driven by citizen and user needs. </em></strong>We use tried and tested methods to understand the lived experience of the people affected by a service or policy, and to learn what it would take to improve.</blockquote><blockquote><strong><em>We know that government is not just transactional.</em></strong><em> </em>We serve the collective interests of the country and recognise that government is about more than individuals wanting simple transactional services. Digital channels are just one way to interact with government.</blockquote><blockquote><strong><em>We are grounded.</em></strong> We are policy engineers, we work on applied policy. We are motivated by practical change, not winning abstract intellectual debates.</blockquote><blockquote><strong><em>We consider the wider system. </em></strong>We understand that making improvements to public services often requires looking beyond the remit of a single team or service. Citizens and service users don’t experience government as an individual piece of legislation, service or department.</blockquote><blockquote><strong>3. Harness the power of boundless horizontal communication</strong></blockquote><blockquote><strong><em>We look outwards. </em></strong>We are unassuming about the role of civil servants in the policy development process and know there is no monopoly on wisdom. We are in touch with the best knowledge outside government.</blockquote><blockquote><strong><em>We are open.</em></strong> Wherever appropriate we are engaged on social media and <a href="https://www.gov.uk/government/publications/social-media-guidance-for-civil-servants/social-media-guidance-for-civil-servants">following guidance</a> we share our working, our challenges and our solutions. We know that <a href="https://www.gov.uk/design-principles#tenth">making things open can make things better</a>.</blockquote><blockquote><strong><em>We collaborate. </em></strong>We draw together information and share experiences and communicate effectively within our teams and professional communities through digital tools and meetups. We use <a href="https://openpolicy.blog.gov.uk/what-is-open-policy-making/">open policy making</a> processes to engage the broadest input possible in the development of content.</blockquote><blockquote><strong><em>We borrow. </em></strong>We build upon the successes of others and don’t start from scratch unless there is no choice.</blockquote><blockquote><strong>4. Embrace the tools and working practices of the internet era</strong></blockquote><blockquote><strong><em>We are as digital at work as we are at home.</em></strong> We make full use of <a href="https://www.gov.uk/government/publications/digital-skills-in-the-civil-service/internet-tools-for-civil-servants-an-introduction">digital tools on the web</a> and we don’t let sub-par government IT, where it still exists, be an excuse to resist reform.</blockquote><blockquote><a href="https://www.gov.uk/design-principles#fourth"><strong><em>We do the hard work to make things simple</em></strong></a>. We take pride in moving from complicated underlying situations to achieving clarity of vision and purpose. We prefer simple language because public services <a href="https://www.gov.uk/design-principles#sixth">are for everyone</a>.</blockquote><blockquote><strong><em>We present clearly.</em></strong> We follow <a href="http://www.doingpresentations.com/">basic rules</a> when we present so we communicate effectively, and wherever possible we <a href="https://gdsengagement.blog.gov.uk/2016/11/04/what-we-mean-when-we-say-show-the-thing/">show the thing</a> we are working on rather than explain it in abstract.</blockquote><blockquote><strong><em>We publish, don’t send. </em></strong>There are often better tools to communicate with colleagues than email.</blockquote><blockquote><strong>5. Use data to inform decisions</strong></blockquote><blockquote><strong><em>We use data to drive both policy and services. </em></strong>We appreciate that analytical insight is as valuable embedded within frontline digital services as it is used in its more familiar setting to inform policy choices.</blockquote><blockquote><strong><em>We are alert to new data techniques.</em></strong> We look to <a href="https://data.blog.gov.uk/category/data-science/">data science</a> to find new insights from data, and we consider the value from non-traditional sources of data. We appreciate but do not solely rely on the traditional analytical professions in government.</blockquote><blockquote><strong><em>We make our data available to others. </em></strong>We know that data can soar in its potential when combined. We publish <a href="https://theodi.org/guides/what-open-data">open data</a> and other content in open formats wherever we can, and make use of authoritative <a href="http://www.openregister.org/">registers</a> to underpin our work. Data security and privacy are critical and we use alternatives to bulk data sharing wherever possible.</blockquote><blockquote><strong><em>We draw on a wide range of analysis: </em></strong>We underpin our work with a variety of analytical techniques, including <a href="https://www.gov.uk/government/organisations/behavioural-insights-team">behavioural science</a>, randomised controlled trials, and evidence from academia and the <a href="https://www.gov.uk/guidance/what-works-network">what works centres</a><strong>.</strong></blockquote><blockquote><strong>6. Pursue an iterative, agile approach to delivery</strong></blockquote><blockquote><strong><em>We start with a presumption to prototype. </em></strong>We believe reforms to most public services can benefit from rapid iteration, grounded in user insight and a relentless focus on rapid, tangible delivery. <a href="http://www.civilserviceworld.com/articles/news/cabinet-office-minister-moots-%E2%80%98maude-award-failure%E2%80%99">When we fail we hope to do it in the right way</a>.</blockquote><blockquote><strong><em>We seek continuous insight.</em></strong> We reject a presumption of multi-year policy cycles, and the notion of a simple linear direction from ministers through policy and then on to delivery. We aim for more than policy work to be just <em>informed</em> by digital delivery, we aim for them to become seamless.</blockquote><blockquote><strong><em>We embrace agile processes to manage our teams and projects.</em></strong> We appreciate practices that reinforce strong team communication and iteration, and not just for digital teams but for any situation <a href="https://twitter.com/psd/status/568324759521501184">where uncertainty is expected</a>.</blockquote><blockquote><strong><em>We get stuff done. </em></strong>We subscribe to the JFDI school of government, over wait and see. We would rather learn by doing than acquire theoretical knowledge.</blockquote><blockquote><strong><em>We are </em></strong><a href="http://oneteamgov.uk/"><strong><em>OneTeamGov.</em></strong></a><strong><em> </em></strong>We reject a hard barrier between our professions. By bringing teams together to work in a truly multidisciplinary way we can offer better services to the public, give better advice to ministers, and make the civil service more effective and efficient.</blockquote><p><strong>A conclusion, for now</strong></p><p>The policy profession still enjoys a privileged position in government at the side of Ministers, but there is a chance that this pre-eminence may come under threat from digital disruption in the coming decade. Rather than leading to a tension or competition between two differing perspectives, the nascent bottom-up <a href="http://oneteamgov.uk/">One Team Government</a> movement in the UK civil service show that there is appetite from many in and around government for these worlds start to flow together in a more seamless way, and in parts even merge.</p><p>The pace of digital change can be unsettling, but it is incumbent upon all civil servants to be better informed about the changes going on, as <a href="https://twitter.com/JanetHughes">Janet Hughes</a> at <a href="https://doteveryone.org.uk/">DotEveryone</a> has said civil servants and others should <a href="https://medium.com/doteveryone/helping-leaders-understand-digital-b2c0e015facc">understand digital technology just like they have to understand money, HR, or the law. Not to be experts — just to have a basic understanding</a>:</p><blockquote><a href="https://medium.com/doteveryone/what-a-digital-organisation-looks-like-82426a210ab8">It is somehow still socially acceptable for leaders to say they don’t understand the changes that are being brought into our lives by digital technology, as though its some kind of niche topic that only specialists need bother themselves with. Digital technology isn’t niche — it affects most aspects of our lives, and most aspects of the strategy and operations of most organisations.</a></blockquote><p>The best way for policy officials to learn to be more digital is to <a href="https://twitter.com/kitterati/status/877234216274202625">just go and work with the large and growing community of digital colleagues</a> in and around government. But there are other resources available to help: for example the <a href="https://www.blog.gov.uk/">100 government blogs</a> and this <a href="https://paulmaltby3.github.io/digitalpolicyreadinglist/Digitalpolicyreadinglist">crowdsourced reading list</a> specifically designed to enable policy people to dig deeper on digital.</p><p>Yet it would be an error for the policy profession to think about digital as just another form of expertise to add to the already long list of competencies that policy officials are expected to be aware of in the course of their duties. Digital is a disruptive technology that is changing the fundamentals of how our society works; <a href="https://www.wsj.com/articles/SB10001424053111903480904576512250915629460">software is eating the world</a>. It would be a category error for policy officials to think about digital as just another consideration at the end of their submission, or as something that is only relevant if proposing a new digital service.</p><p>The policy world has been moving towards increasing expertise for many years. Elements that used to be part of the generalist’s role, such as operational delivery, communications, programme management and commercial expertise have become professionalised. Rather than undermine or weaken the policy profession, in this further moment of change we are faced with an opportunity to better clarify and celebrate the core policy skill set — and for all its disruption the digital community is in deep need of these skills.</p><p>Saying we know what policy is when we see it just isn’t enough any more for the credibility and self-respect of policy officials. We should be embracing the digital side of policy and offering the policy profession’s unique insights into these new systems and ways of working from a position of strength. It is time for us to move more confidently out from the shadows (where even our basic ‘how to’ guides on policy are hidden behind <a href="https://civilservicelearning.civilservice.gov.uk/professions/professions/policy/about-policy-profession">password protected firewalls</a>) to create a more open, forward looking and collaborative culture and set of skills. Engaging in the <a href="http://oneteamgov.uk/">One Team Government</a> community and events is just the start.</p><p><em>Follow One Team Government on </em><a href="https://twitter.com/OneTeamGov"><em>Twitter</em></a><em> or via the </em><a href="http://www.oneteamgov.uk/"><em>website</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1596d99d3f57" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[One Team Government: What digital and policy can learn from each other]]></title>
            <link>https://medium.com/@maltbyps/one-team-government-what-digital-and-policy-can-learn-from-each-other-c284b31b75ac?source=rss-d3a206a02212------2</link>
            <guid isPermaLink="false">https://medium.com/p/c284b31b75ac</guid>
            <category><![CDATA[policy]]></category>
            <category><![CDATA[government]]></category>
            <category><![CDATA[innovation]]></category>
            <category><![CDATA[digital]]></category>
            <dc:creator><![CDATA[Paul Maltby]]></dc:creator>
            <pubDate>Wed, 28 Jun 2017 07:26:28 GMT</pubDate>
            <atom:updated>2017-09-15T09:25:56.169Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/912/1*PNyIxnx7GDoHHfuUkLyvRg.png" /></figure><p><em>This is the third of four posts published in the run up to the </em><a href="http://oneteamgov.uk/"><em>One Team Government</em></a><em> unconference on 29 June 2017, where a bottom-up community of digital and policy officials in and around the UK government are coming together to learn more about each other’s work and to explore how these roles are starting to merge.</em></p><p><em>The first post was a </em><a href="https://github.com/paulmaltby3/digitalpolicyreadinglist"><em>crowdsourced reading list </em></a><em>for policy officials interested in digging deeper on digital, which is posted on GitHub to allow updates and new recommendations.</em></p><p><em>The second was a </em><a href="https://medium.com/@maltbyps/one-team-government-a-short-guide-to-policy-for-government-digital-professionals-c3cc1e421406"><em>guide to policy for digital professionals</em></a><em> in and around government.</em></p><p><em>A subsequent post will propose a prototype for how digital policy officials in government might work — </em><a href="https://medium.com/@maltbyps/prototyping-a-one-team-government-manifesto-1596d99d3f57"><em>a One Team Government manifesto </em></a><em>— which is designed to be subject of a fuller discussion at the unconference on 29 June</em></p><p><em>Drawing on my experience working across the space between policy and digital (and data) in government, this post reflects on what each of these professions might best learn from the other.</em></p><p><strong>1. What policy can learn from digital</strong></p><p>The last few years have seen a rise in<a href="https://www.gov.uk/government/organisations/digital-data-and-technology-professions"> digital professions</a> and associated ways of working within government, encompassing not just technical roles but others in wider <a href="https://www.gov.uk/service-manual/the-team/what-each-role-does-in-service-team">agile delivery teams</a>, including service designers, product managers, user researchers and more. While technology is of course an important factor, it is just one of the elements that have changed — a point well made in Tom Loosemore’s definition:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/625/1*q0P7y-z6T3QyRi2Z51FQBA.png" /></figure><p>The policy profession has not been immune from criticism over the years, but the rapid emergence of digital and its transformative impact both inside and outside government has highlighted new challenges alongside more familiar complaints.</p><p><strong><em>‘The policy process is too rigid and linear’</em></strong></p><p>One of the unstated myths of policy is that there is a <em>presumed directional hierarchy</em><strong><em> </em></strong>at work in civil service activity: a process that starts with a vision or particular details from ministers, refined through options and analysis from policy officials, with decisions then flowing on through into a handover to those in operational professions who implement those decisions. Although the process is often formally presented as a circular process (eg the HM Treasury Green Book ‘<a href="https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/220541/green_book_complete.pdf">ROAMEF’</a>), the circular link in the chain has always been understated, and any cycle is presumed to operate over a period of years or even Parliaments.</p><p>It is the presumption of a linear, insular policy process that is perhaps most frustrating and alien to government digital teams. For digital teams the break from ‘waterfall’ project management to<a href="https://www.gov.uk/service-manual/agile-delivery/agile-government-services-introduction"> Agile processes</a> with its rapid iteration of tangible products based on carefully assessed user need is not just about building effective digital services, it is an acknowledgement of the powerful cultural and behavioural shifts that have come about in the modern digital era. Policy’s frequent ignorance at the seismic cultural and practical shifts created as a result of the<a href="http://agilemanifesto.org/"> Agile Manifesto</a> makes the profession look like it has not moved with the times when it interacts with digital, even though good quality policy has always involved flexibility, iteration and a deep understanding of its use and impact in the real world.</p><p>The digital world often moves with incredible speed as it develops services, with tangible prototypes and <a href="https://www.gov.uk/service-manual/agile-delivery/how-the-alpha-phase-works">alpha</a> versions of products sometimes produced within days, not years. In the the meetings-based culture of policy it is not unknown that during the period waiting for diaries to become free to discuss the basic premise of the service, the digital team have already produced a basic product. The contrast can be infuriating.</p><p><strong><em>‘There is a gap between policy and users/delivery’</em></strong></p><p>The gap between policy and operations has long been criticised by Ministers and senior officials alike (“<a href="https://www.instituteforgovernment.org.uk/sites/default/files/publications/Policy%20making%20in%20the%20real%20world.pdf">very, very poor connections between policy advice and delivery experience, so often policy ideas are put forward in the absence of any understanding whatsoever in how you actually deliver anything</a>”), and the privileged position of policy ‘above’ other disciplines in government breeds a common criticism that policy would rather engage in the intellectual pursuit of ideas and writing rather than the ‘dull’ <a href="https://www.instituteforgovernment.org.uk/blog/john-manzoni-sets-out-his-priorities-chief-executive-civil-service">functional agenda</a> of managing people, buildings and contracts. There is perhaps not a little irony that as policy officials specialise and develop skills in other areas there is a fear among the profession — <a href="http://data.parliament.uk/writtenevidence/committeeevidence.svc/evidencedocument/public-administration-and-constitutional-affairs-committee/the-work-of-the-civil-service/written/47070.html">and some Ministers</a> — that they are losing their core ability to reason, write and craft a quality briefing note and communicate in a sophisticated way with Ministers.</p><p>There is also a common complaint from digital that policy can appear to lack <em>clarity or sometimes even an interest in its end-users </em>and what drives their experience of a situation or service. Good policy makers will always have gone out and experienced operational delivery of services and talked with and observed citizens interact with services, and there is an important challenge to digital’s presumption that government is built on transactional services and individualistic relations between government and users, and this is addressed in the section below. But too often policy’s engagement with users is not done frequently or routinely, and it is almost never done with the <a href="https://www.gov.uk/service-manual/user-research">serious intent and methodology</a> seen in the practice of designers and specialist user-researchers. Nor is there anything like the basic presumption and hunger for end-user feedback seen in digital working practice. When introduced to these techniques through the work of <a href="https://openpolicy.blog.gov.uk/">Policy Lab</a> or through close working with digital colleagues it can be surprising how intense the positive reaction from policy officials can be, and how shocked colleagues are that have not embraced these quite basic and widely used design techniques before.</p><p>From a digital point of view the central focus on user needs leads to a common desire to turn the traditional policy making process ‘on its head’, as Mike Bracken said his <a href="http://mikebracken.com/blog/on-policy-and-delivery/">critique of policy making in government</a>, and to radically improve its <em>pace and rhythm</em>. Compared with the rapid iteration of digital products, policy is slow and inflexible, and at times it can appear to fetishise complexity rather than <a href="https://www.gov.uk/design-principles#fourth">reward the hard work it takes to make thing simple</a> — a deeply felt principle among design professionals. The thinking goes: if user needs have identified an unintentional policy or legislative barrier to improving a service, how can it take such a long process of green and white papers, formal consultations and possibly legislation to make a change? Even if some parts of the process like legislation are unlikely to speed up any time soon we should be able to develop processes that start with a Ministerial mandate and which, through prototyping, flow rapidly around the cycle as services iterate, and flow in all directions. We should start with the expectation that policy changes will come flowing upwards as digital service delivery interacts with users, and policy decisions should routinely be seen as moments that establish a prototype and are merely the start of an iterated process, not the final end-point of service design.</p><p><strong><em>‘Policy is blind to its (dated) assumptions’</em></strong></p><p>All organisations carry with them their own culture and assumptions about how best to intervene, and the civil service in general and departments specifically are not immune from this. Some departments tend to reach for legislative solutions to new challenges, others to a new programmes and a ring-fenced pot of money. While in theory we bring the right tool to each job, in practice there are often unconscious assumptions in play about what sorts of activity are deemed appropriate.</p><p>Digital teams are often fresh faces in a department and bring a different organisational culture. They can be quick to spot these cultural blindspots at work in a department, and not just those related to a lack of technical understanding and digital ways of working. Calling out these assumptions can feel threatening to others in a department, even when done with benign intentions. At times policy and other more established civil service professions can be reluctant to rock the boat and can value process over outcomes, complexity over clarity, general knowledge over deep technical skill. Phrases like ‘we tried that before and it didn’t work’ delivered with a roll of the eyes can sap the energy of reformers in government regardless of their profession. And generalists too often act as an intermediary between those with specialist skills and decision makers — we need to quickly move to a position that individuals with deep digital delivery skills are present in senior decision making meetings with Ministers or senior officials.</p><p>The assumption that policy work has to mean closed working in private can be particularly testing for digital practitioners, where the shift towards <a href="https://en.wikipedia.org/wiki/The_Cathedral_and_the_Bazaar">open source development happened a couple of decades</a> ago and where collaborative tools like <a href="https://stackoverflow.com/">StackOverflow</a> and <a href="https://github.com/">GitHub</a> are near ubiquitous. The GDS design principle of ‘<a href="https://www.gov.uk/design-principles#tenth">make things open it makes them better</a>’ is not just practical advice, it reflects a broader cultural approach and set of implicit behaviours. A combination of the UK’s adversarial political system, the ultimate accountability of Ministers to Parliament, and a healthy desire to <a href="https://ico.org.uk/media/for-organisations/documents/1200/government-policy-foi-section-35-guidance.pdf">protect safe spaces around Ministers to enable the debate of controversial options</a> means that there has been a reluctance for policy officials to engage too broadly or visibly in policy debates. On controversial political issues this can be sensible — the<a href="https://www.gov.uk/government/publications/social-media-guidance-for-civil-servants/social-media-guidance-for-civil-servants"> social media guidance</a> from Cabinet Office remains a good guide to navigating this sometimes difficult territory.</p><p>But the assumptions about working on policy only in private are being challenged, not least through the movement towards open policy making and with leadership from <a href="https://openpolicy.blog.gov.uk/2014/09/22/heywood-video-1/">the Cabinet Secretary Jeremy Heywood</a> who has challenged the presumption that policy officials have a ‘monopoly on wisdom’, and from <a href="https://twitter.com/PolProfHeadUK">Chris Wormald the head of the policy profession</a> in government and Permanent Secretary who when at the Department for Education introduced <a href="https://quarterly.blog.gov.uk/2013/07/12/the-policy-tests-transforming-policy-in-the-department-for-education/">five policy tests</a> for officials that included the provocative question ‘who made you the expert?’.</p><p>A more collaborative policy mindset does not have to mean doing everything in the open, and there are good examples from departments like the <a href="https://quarterly.blog.gov.uk/2013/10/22/reinvesting-in-policy-making/">Foreign and Commonwealth Office</a> where security and discretion are critical. But old habits persist, occasionally with strong Ministerial support. In general there remains a gulf between the type of open problem solving seen from mainly digital teams in around <a href="https://www.blog.gov.uk/">100 government blogs</a> and the day to day activity of most policy civil servants. For digital teams this not only looks like an inefficient way to solve problems but it looks arrogant and insular, with a suspicion that information is withheld as a form of control and power. And it can make policy officials look like they are hiding from view, <a href="http://www.hollidazed.co.uk/2017/06/27/work-in-the-open-and-finding-your-edge/">wary of personal accountability and responsibility</a>.</p><p><strong>2. What digital can learn from policy</strong></p><p>Digital disruption is not a straightforward incremental process and new paradigms can mean leaving some sensible ways of working behind. So be it. This blog post doesn’t argue for the policy/digital equivalent of <a href="https://en.wikipedia.org/wiki/Wagile_software_development">‘wagile’</a> (the often <a href="https://gds.blog.gov.uk/2015/07/10/you-cant-be-half-agile/">unfortunate marriage</a> of waterfall and agile project management approaches). Half reforming is a strange ambition.</p><p>That said, while government is spending serious effort getting the mainstream civil service to understand digital there seems much less out there on what the digital professions can learn from policy, and for those of us that straddle these two professions it feels like there are some blind spots that need recognising.</p><p><strong><em>‘Government is not all transactional’</em></strong></p><p>One of the most perplexing features for policy officials working with some digital colleagues is the overwhelming focus on the transactional services government delivers, compared to more collective government activity.</p><p>Transactions are where an individual wants something from government and for example goes online to sort it out. The challenge to <a href="https://gds.blog.gov.uk/2013/07/17/the-pivot-from-publishing-to-transactions/">fix transactions</a> led to GDS’ <a href="https://www.gov.uk/transformation/exemplars">Transformation Programme</a> of 25 exemplar digital services in 2013. The result has been a massive uplift in the citizen’s experience of these services alongside a significant reduction in costs. It needed fixing, and the experience of many digital colleagues who have similarly revolutionised improvements to services in the private sector has been invaluable.</p><p>But government is as much about ‘we’ as ‘me’. The basic purpose of state intervention is <a href="https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/220541/green_book_complete.pdf">‘usually founded either in market failure or where there are clear government distributional objectives that need to be met</a>’, often in instances where the issue in question raises societal interests beyond an individual. For example, the policy choices around the programme of <a href="https://www.gov.uk/community-sentences/treatment-and-programmes">interventions</a> for individuals convicted of crimes to prevent reoffending is a mixture of interests of the individual in question, their family and friends, neighbours, and the taxpayer. Similar issues are in play when working out whether to embark on a major infrastructure project like <a href="https://www.gov.uk/government/organisations/high-speed-two-limited">HS2</a>, and if so where to locate it; deciding whether accountability regimes in policing, education or health are appropriate and whether change could improve outcomes for citizens; and dealing with so-called <a href="https://www.gov.uk/government/news/troubled-families-programme-annual-report-published">troubled families</a>. There seem few big government decisions that don’t involve positive or negative <a href="https://en.wikipedia.org/wiki/Externality">externalities</a> of one type or another. Working out how to identify and then manage tensions between competing interests in these situations is a mainstay of policy work in government.</p><p>It is not just the gravitational pull towards transactions and away from more messy government activity that can be a frustration for policy officials. It is that digital teams who are grounded in engineering products can make policy decisions on the fly, sometimes without realising, and often without being familiar with the basic policy processes to work through different options. Decisions that can just look sensible or efficient can miss bigger policy and political issues, especially when they have repercussions on citizens outside a narrow definition of ‘user’, when speed of delivery is paramount, and when digital teams can at times have an overly homogenous outlook on life. Learning how to identify hidden ideologies located within product choices and dealing with competing possibilities and ideas is perhaps one of the remaining benefits of a liberal arts degree in a digital government world.</p><p>There is also an implication for our adherence to the centrality of user needs when the collective purpose of government is acknowledged. The digital revolution in services and products in our wider economy has been founded on a relentless focus on user needs, and in government the first digital design principle has acted as a critical call to action against self-regarding producer interest and civil service insularity: ‘<a href="https://www.flickr.com/photos/dafyddbach/9220817053/">user needs not government needs</a>’. But this presupposes a binary between individual recipients of services and a self-serving bureaucracy, and it does not make space for recognising <em>collective</em> citizen interests that flow from externalities. Ultimately, it may be that user needs are necessary but not sufficient within a wider government context, and that ‘citizen needs’ could be a better focal point.</p><p>There is a risk that an overly narrow focus on the user in situations where there are contested moral, political and social choices (divorce, benefits, criminality etc) can lead to naive conclusions. It is a paradox of modern digital government that despite the deep civic mindedness of many of the individuals involved, the individualistic, transactional mindset that drives the digital economy can at times seem to override a more collectivist, social perspective. Collective citizen concerns are not an <a href="https://en.wikipedia.org/wiki/Edge_case">edge case</a>, and a perspective that is reluctant to put an individual citizen and their individual control at the centre of a digital world is not always a sign of a bureaucracy interested in furthering its own agenda or failing to appreciate the significant insights from user needs analysis.</p><p><strong><em>‘Democracy &gt; user needs’</em></strong></p><p>Policy professionals also have a range of other skills highlighted in the <a href="https://medium.com/@maltbyps/one-team-government-a-short-guide-to-policy-for-government-digital-professionals-c3cc1e421406">previous post</a> that can be invaluable in making progress in government, and which are sometimes lacking in digital teams. For example, working through competing upstream options, judgement in intervening in political or contested environments, managing ambiguity and exercising nuance in dealing with organisations and people, and — occasionally — busting through hype and technological solutionism with a healthy dose of scepticism.</p><p>There is also an important but frequently misunderstood skill in creating the right amount of distance between an individual and the content they are working on. While there are few things more frustrating than the cold aloof civil servant seemingly indifferent as to whether the issue they are working on succeeds or fails, there is something to be said for striking an appropriate balance and drawing on a realist and sometimes sceptical approach. The civil service serves whichever Ministers the public choose in general elections, and one of the unusual features of the job is that this can mean having to pivot in the opposite direction with the same degree of enthusiasm and commitment — a psychologically odd process that is mitigated through our strange ritual of preparing <a href="https://www.civilserviceworld.com/articles/opinion/general-election-2017-preparing-day-one-new-minister">first day briefing packs</a> for new ministers.</p><p>Of course, not everything changes overnight with the appointment of new ministers or a new government, and this is usually more true the closer one gets to operational delivery roles. But there is something about the infectious passion of digital colleagues who are newer to government that can at times start to feel like the all-conquering belief in user needs is being mistaken as a ‘factual’ proof of what has to change. Democracy is our best corrective process to manage the full range of views, emotions, needs and hopes of the country; and it is profoundly bigger than user needs. Many of us fight daily to improve the process of government so that it can speedily and accurately bring the user perspective close to Ministers and other decision makers in government. But sometimes Ministers will make a decision that does not seem to reflect this ‘rational’ insight established from user needs, and that is absolutely fine. All civil servants — digital practitioners included — need to be utterly committed to implementing those decisions. A number of us have too often witnessed times when ‘user needs’ that coincidentally reflect the views of the individual in question have been used as an excuse for why a thing can’t or must happen. Ministers understandably have a deep antipathy to civil service interests trumping legitimate democratic decisions, even (especially?) when these are framed as empirical knowledge of what citizens need.</p><p><em>The next post in this series will look at </em><a href="https://medium.com/@maltbyps/prototyping-a-one-team-government-manifesto-1596d99d3f57"><em>how we can bring policy and digital together in a more seamless way of working</em></a><em> and develop a prototype for a One Team Government manifesto.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c284b31b75ac" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[One Team Government: a short guide to policy for government digital professionals]]></title>
            <link>https://medium.com/@maltbyps/one-team-government-a-short-guide-to-policy-for-government-digital-professionals-c3cc1e421406?source=rss-d3a206a02212------2</link>
            <guid isPermaLink="false">https://medium.com/p/c3cc1e421406</guid>
            <category><![CDATA[innovation]]></category>
            <category><![CDATA[oneteamgov]]></category>
            <category><![CDATA[policy]]></category>
            <category><![CDATA[digital]]></category>
            <category><![CDATA[government]]></category>
            <dc:creator><![CDATA[Paul Maltby]]></dc:creator>
            <pubDate>Tue, 27 Jun 2017 07:34:55 GMT</pubDate>
            <atom:updated>2017-06-29T05:34:14.067Z</atom:updated>
            <content:encoded><![CDATA[<p><em>This is the second of four posts I’m publishing in the run up to the </em><a href="http://oneteamgov.uk/"><em>One Team Government </em></a><em>unconference on 29 June 2017. One Team Government is a new bottom-up community of digital and policy officials in and around the UK government who are coming together to learn more about each other’s work and to explore how these roles are starting to merge.</em></p><p><em>My first post was a </em><a href="https://github.com/paulmaltby3/digitalpolicyreadinglist"><em>crowdsourced reading list</em></a><em> for policy officials interested in digging deeper on digital, which is posted on GitHub to allow updates and new recommendations.</em></p><p><em>Two subsequent posts will discuss </em><a href="https://medium.com/@maltbyps/one-team-government-what-digital-and-policy-can-learn-from-each-other-c284b31b75ac"><em>what the digital and policy professions might best learn from each other</em></a><em>, and propose a prototype for how digital policy officials in government might work — a </em><a href="https://medium.com/@maltbyps/prototyping-a-one-team-government-manifesto-1596d99d3f57"><em>One Team Government manifesto </em></a><em>— which is designed to be discussed at the unconference on 29 June</em></p><p><strong>Why this post?</strong></p><p>In Whitehall ‘policy’ and ‘digital’ can at times feel like they are entirely different universes, populated by colleagues with significantly different skill sets. But these worlds are colliding, and the working practices and underlying assumptions of both disciplines are likely to be challenged and perhaps significantly so as the digital revolution continues its inexorable pace. The two disciplines have<a href="https://www.instituteforgovernment.org.uk/events/digital-government-strategy-delivery-mike-bracken-head-government-digital-service-1"> not always seen eye to eye </a>as digital has wrestled for its place at the most senior Whitehall tables, but there is a new cadre of officials coming together under the banner of <a href="http://oneteamgov.uk/">One Team Government</a> that are rewriting the way in which government works and are curious about each other’s disciplines.</p><p>This article is written from this same spirit of positive disruption. There has been a variety of work across government to explain the new world of digital to existing civil servants, but perhaps less attention has been given to explain the world of policy to digital colleagues, and even less has been done to figure how these two disciplines need to entwine in our new world. This article has come from having had a foot in both policy and digital/data in the Cabinet Office and Government Digital Service in recent years after a longer period as a public service reformer and policy official within big Whitehall departments and local government. It seeks to answer a question often posed quietly by individuals in digital teams: what <em>is</em> policy?</p><p><strong>What <em>is</em> policy?</strong></p><p>Policy is a word used to describe the government’s position or approach to a particular issue. The term ‘policy profession’ has come into use <a href="https://civilservicelearning.civilservice.gov.uk/sites/default/files/twelve_actions_report_web_accessible.pdf">in recent years</a> to describe those civil servants who advise and carry out this work, having emerged from the historical and increasingly inadequate catch-all description of the civil service ‘generalist’.</p><p>Despite policy’s presumed pre-eminent position around government, as a newly established profession we have sometimes struggled to describe what it is that we do. In guides to policy making for new civil servants we introduce our profession using jargon (“delivering change in the real world”), or in tautological terms (“when we talk about policy we are talking about the decisions made by the government in light of the advice provided by civil service policy officials”).</p><p>Perhaps a more specific description is that policy work is about <em>drawing on familiarity with evidence, politics and delivery to advise Ministers about an appropriate course of action, and then putting in place activity to make it happen.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/427/0*DKr8AZw9Ru8oG54l." /></figure><p><em>Figure: ‘Three key elements in policy making’, </em><a href="https://civilservicelearning.civilservice.gov.uk/sites/default/files/twelve_actions_report_web_accessible.pdf"><em>Twelve Actions to Professionalise Policy Making</em></a><em>, 2013</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/870/0*uNpcoqA160nX_NLf." /></figure><p><em>Figure: </em><a href="https://civilservicelearning.civilservice.gov.uk/sites/default/files/2016-12-21_policy_profession_standards_v3.3.pdf"><em>Policy Profession Standards Framework</em></a><em>, 2017</em></p><p>The profession expects that civil service policy officials will have an intellectual grasp of the possibility of different political and practical interpretations and interventions that cover not just the ‘the what’ but also ‘the how’, and will use these abilities to challenge assumptions and evaluate different potential options. Officials are expected to have the ability to understand not just the detailed matter in question, but also a much wider context comprising of politics, citizen pressure, media, business and other stakeholder views and how all these factors could influence and be influenced by different recommended actions. As a result, regardless of the narrowness of the issue in question, policy officials are alert to the potential for most government activity to impact on a much wider public arena and also to have an impact on the collective experience of the public realm, and not just individual citizens experiences of government.</p><p>Policy professionals are expected to have an understanding of when, why and how government might intervene on a particular issue (see <a href="https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/220541/green_book_complete.pdf">HM Treasury Green Book p11</a>), and be especially alert to different types of market failure and (eg public goods, information asymmetries, monopoly) or matters of equity or security. They are also expected to be conscious of different forms of intervention (including the possibility for no intervention), including:</p><ul><li>Legislation, regulations or other Parliamentary or legal intervention</li><li>Funding a type of activity, programme or commissioning external activity</li><li>Designing or amending a service or intervention, including establishing pilots, prototypes or demonstration projects</li><li>Choosing the appropriate structures through which a service should be administered, or a system given oversight: for example in-house provision, arms-length body, commissioned charity or social enterprise, commercial provider, or through a free market regulated to different degrees</li><li>Intervening with or galvanising an external community or profession, running campaigns aimed at particular professions or a wider group of citizens</li><li>Establishing standards, principles or voluntary codes of behaviour</li><li>Setting the tone and parameters for discussion, drawing attention to issues, interfacing with media and communications, communicating narratives</li><li>Analytical and innovative intervention techniques, including behavioural science, data science, use of social media and other digital techniques</li></ul><p>The skills required of policy officials include an ability to navigate and cut through complex bureaucracies and to be aware of how power and authority are transmitted through organisations, cultures and people, how incentives impact on them, and assessing how much change and of what type the situation will bear. They are rewarded for clear thinking and intellectual and ethical rigour and challenge, alongside a multi-faceted knowledge comprising of wide range of skills: intellectual, analytical, political awareness and history, legal, communications, commercial, economic, financial, project management, statistical/numerate, digital, and operational disciplines.</p><p>Policy officials are expected to develop and retain deep domain expertise and situational awareness on a topic, and to be able to communicate this knowledge clearly and succinctly in written and verbal formats, and where necessary persuade and argue a point successfully — regardless of their own personal point of view. The best have a good dose of creativity in problem solving which is used to tackle a challenging situation, and the ability to make deals and understand the commercial realities of a situation are increasingly prized. Alongside all this, policy officials have always been expected to be action-focussed and to understand the constraints and opportunities on policy from delivery and operations, although it is frequently these points that can be a point of contention in discussions about the interface with policy and digital (and a subject of the next post)</p><p>Also, ‘policy’ is not a single type of activity: it varies widely along what is sometimes presented as a spectrum from strategic policy covering the broadest questions (eg ‘what should the accountability regime for schools look like’), through the design of a particular public service in response to Ministerial or legal challenge, to much more specific rule setting for operational officials (eg ‘what are the operational procedures that must be followed by officials at the border as a result of the Immigration Act and subsequent court rulings’). It sometimes involves work directly with Ministers (eg in private office), or in Parliament (managing a particular piece of legislation), but routinely it involves being responsible as a general <a href="https://www.instituteforgovernment.org.uk/sites/default/files/publications/System%20Stewardship.pdf">system steward</a> for a particular area of government activity and response (eg planning policy within a Housing and Planning Directorate), perhaps with responsibility for particular arms-length bodies, sections of industry, and sometimes also ownership of front-line services.</p><p>Successful policy officials come in many different guises, but telling phrases that resonate include the ability to ‘manage ambiguity’, expected in the relatively common situation when different outcomes or mechanisms are desired from the same policy area by senior figures of equal power and influence, and yet for reasons outside the officials’ control the ability to bring about a resolution to that ambiguity is not possible. A counterpoint is the emphasis put on the term ‘judgement’, which can involve an official’s ability to accurately assess the situation in an especially complex matter that requires knowledge across different disciplines, and to find a route to cut through and bring about progress on that topic.</p><p><em>The next post in this series will look at </em><a href="https://medium.com/@maltbyps/one-team-government-what-digital-and-policy-can-learn-from-each-other-c284b31b75ac"><em>what the policy and digital professions in government might learn from each other</em></a><em>.</em></p><p><strong>More reading</strong></p><p><a href="https://civilservicelearning.civilservice.gov.uk/sites/default/files/twelve_actions_report_web_accessible.pdf">Twelve Actions to Professionalise Policy Making</a></p><p><a href="https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/220541/green_book_complete.pdf">HM Treasury Green Book</a></p><p><a href="https://www.instituteforgovernment.org.uk/sites/default/files/publications/Policy%20making%20in%20the%20real%20world.pdf">Institute for Government, Policy Making in the Real World</a></p><p><a href="http://openpolicy.blog.gov.uk/">Policy Lab and Open Policy Making blog</a></p><p><a href="https://quarterly.blog.gov.uk/">Civil Service Quarterly</a></p><p><a href="https://www.gov.uk/government/organisations/civil-service-policy-profession/about">The Policy Profession</a> on GOV.UK</p><p><a href="https://www.gov.uk/guidance/what-works-network">The What Works Network</a></p><p><a href="https://civilservicelearning.civilservice.gov.uk/professions/policy-profession">Civil Service Learning Policy Profession</a> (requires government email/password)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c3cc1e421406" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>