<?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 Livnat Peer on Medium]]></title>
        <description><![CDATA[Stories by Livnat Peer on Medium]]></description>
        <link>https://medium.com/@livnats?source=rss-7174840c036e------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*PK4Zz0Bv2xgtJVqaMnA0Xw.jpeg</url>
            <title>Stories by Livnat Peer on Medium</title>
            <link>https://medium.com/@livnats?source=rss-7174840c036e------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 08:38:52 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@livnats/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[How AI Will Disrupt Everything You Know About DevOps]]></title>
            <link>https://livnats.medium.com/how-ai-will-disrupt-everything-you-know-about-devops-502a2ece61b1?source=rss-7174840c036e------2</link>
            <guid isPermaLink="false">https://medium.com/p/502a2ece61b1</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[agents]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <category><![CDATA[platform-engineering]]></category>
            <dc:creator><![CDATA[Livnat Peer]]></dc:creator>
            <pubDate>Sun, 04 May 2025 13:30:15 GMT</pubDate>
            <atom:updated>2025-05-04T13:30:15.600Z</atom:updated>
            <content:encoded><![CDATA[<p>This year, I had the honor of delivering the keynote at the <a href="https://www.leaders-in.tech/">“Leaders in Tech”</a> conference, opening <a href="https://www.leaders-in.tech/agenda/">DevOps Day</a>. I spoke about how AI is set to disrupt everything we know about DevOps — in this post, I’d like to share the key ideas from the session.</p><p>Today’s DevOps landscape has come a long way from the days of manual testing, midnight deploys, and endless war rooms. The DevOps revolution, nearly a decade ago, promised faster delivery, infinite scale, and seamless collaboration — and it delivered. A lot.</p><p>While DevOps has undeniably pushed us forward, bridging the gap between development and operations, many of its original promises remain aspirational rather than reality. Pipelines still break, alerts still flood in (most of them noise), and teams still argue over who’s to blame. We’ve optimized everything we can control, yet the cracks are showing. Why? Because complexity is exploding. The number of systems we manage keeps growing, and it’s getting harder to keep up.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*dPtmnjmP53WgkH6f" /><figcaption>Devops — The cracks are showing</figcaption></figure><p>This isn’t failure — it’s a signal. The next leap won’t come from more tools or better YAML. It’s going to come from AI.</p><h4><strong>DevOps Reimagined</strong></h4><p>In a future where the GenAI revolution has fully matured, DevOps becomes a self-sustaining, hyper-intelligent ecosystem.<br>Code isn’t written line-by-line — AI agents translate business goals into production-ready systems, architecting infrastructure, writing secure code, and constantly optimizing performance.<br>Monitoring systems, powered by predictive AI, anticipate outages before they occur, spinning up fixes or rerouting traffic autonomously.<br>Alerts are no longer noise — they’re rare, meaningful, and often resolved before anyone even notices.<br>Incident response evolves from reactive firefighting to proactive system refinement, as GenAI learns from anomalies and continuously improves.<br>Human–machine collaboration becomes seamless, with engineers guiding a symphony of intelligent agents that build, scale, and secure systems with speed and precision.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*q4nPZm1mwzNtkUf1" /><figcaption>Devops — ReImagined</figcaption></figure><p>This might sound futuristic — even imaginary — but the shift has already begun. The technology is here, it’s up to us to explore what’s possible and start shaping what comes next.</p><h4><strong>The Journey</strong></h4><p>The journey from today’s DevOps to a future shaped by a self-sustaining, hyper-intelligent AI ecosystem is one of transformation — and it starts with a mindset shift. We need to be open, adaptable, and ready to embrace change at every turn. This evolution isn’t a single leap, but a series of intentional stops, each representing meaningful progress.</p><p>Start with the mindset shift, many people are asking real, hard questions: ‘What does this mean for my job?’, ‘Will I be replaced?’, ‘Where do I fit in a world run by digital agents and AI tools?’. These aren’t just hypotheticals — they’re lived concerns, especially in industries already feeling the ripple effects.<br>Empathy has to lead this conversation. Because the truth is, we don’t have all the answers. And pretending we do would only deepen the divide between the technology and the people it’s meant to serve.</p><p>But here’s what we do know: the technology is <em>not waiting</em> for us to feel ready. It’s here, and it’s accelerating. Those who resist this shift risk being left behind — not because they’re not smart or capable, but because the rules of relevance are changing. <em>Fast</em>.</p><h4>Plan your Strategy</h4><p>A meaningful step in this Journey is planning your AI strategy. Your strategy should include three pillars: Tools echosystem, Digital Workforce and Experimentation.<br><strong>Tools echosystem</strong> — Start with making deliberate choices — don’t just turn on AI everywhere and hope for magic. Choose the tools that are <em>strategic</em> to your workflows and business goals. Enabling AI features across every system might sound impressive, but it’s expensive, duplicative, and often leads to more chaos than clarity. Focus your efforts where they matter most.<br>From there, think about how you build your <strong>Digital Workforce</strong> — your AI -powered teammates. Digital Workforce represents a strategic deployment of AI agents designed to augment human capabilities, accelerate workflows, and scale operations across the enterprise.<br>Digital agents are software programs designed to autonomously execute tasks. Start small. Begin with simple, measurable tasks that solve real problems, and evolve your agents as your understanding and confidence grow. <br>Finally, leave space to <strong>Experiment</strong>. AI is moving fast, and what works today may not be best tomorrow. Keep testing, iterating, and learning to stay ahead, not just afloat.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8kQ6e4iAbjAHRvBQ6UtLnw.png" /><figcaption>3 pillars in your AI Strategy</figcaption></figure><h4><strong>Digital Workforce</strong></h4><p>Once you’ve defined a solid strategy, it’s time to think about your digital workforce — intelligent agents you can start using today with existing technology. <br>When thinking about digital workforce for DevOps, a helpful mental model is the classic DevOps infinite loop — plan, develop, build, test, release, deploy, operate, and monitor. Each phase of this loop represents an opportunity to embed intelligent agents that can augment human efforts, automate routine tasks, and enhance decision-making. By mapping AI agents to each stage, we can start to imagine — and implement — a smarter, more adaptive DevOps workflow.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/952/1*B4QroDDJ6gZdRoxcNn_4zA.png" /><figcaption>DevOps infinate loop</figcaption></figure><p>One of the easiest wins? A <strong>code reviewer agent</strong>. Tools like <a href="https://www.coderabbit.ai/">CodeRabbit</a> or <a href="https://sourcery.ai/">Sorcery AI</a> can act as automated reviewers, scanning every commit for issues, suggestions, and improvements. If your current workflow requires two human reviews before merging, you can shift to one human review and one AI review — maintaining rigor while improving speed and consistency.<br>Another powerful addition is a <strong>code generator agent</strong> — an AI agent that can generate code snippets, Ansible playbooks, Docker files etc. based on natural language prompts or context within your project. The most effective setup is one that integrates directly into your IDE, bringing the agent into your day-to-day workflow. Tools like <a href="https://github.com/features/copilot">GitHub Copilot</a>, <a href="https://www.cursor.com/">Cursor</a>, or <a href="https://windsurf.com/editor">Windsurf</a> are very useful.<br>In the testing domain, a <strong>test agent</strong> goes beyond just generating test code — though it should certainly do that, much like a code generator. Additional capabilities include intelligently selecting and running only the most relevant test suites for each commit, significantly reducing CI/CD runtimes and saving compute resources. Test agents can also generate realistic mock data on the fly, making it easier to validate business use cases.<br>The <strong>incident manager agent</strong> is your AI teammate during high-pressure moments. It can summarize incidents in real time, reducing handover friction and quickly bringing the right people into the loop. By pulling logs from multiple systems and enriching them with contextual data, it builds a clearer, more actionable picture of what’s happening. It also documents every step of the incident response as it unfolds — capturing decisions, actions, and outcomes for post-incident reviews and continuous improvement.<br>These agents aren’t science fiction, they are just a few example to a large variety of agents that should be integrated into your workflows. They do not cover all the stages in the devops loop, but are examples to what you can do. By thoughtfully adopting AI-powered teammates like code reviewers, generators, testers, and incident managers, you’re not just automating tasks — you’re evolving your DevOps culture toward greater speed, resilience, and intelligence.</p><h4>The Challenges</h4><p>As we move toward a GenAI-driven DevOps future, it’s crucial to recognize the challenges along the way. GenAI models thrive on large volumes of high-quality data — but most enterprise IT environments aren’t ready. As <a href="https://www.linkedin.com/posts/armand-ruiz_the-operating-principles-of-enterprise-ai-activity-7317502406575984640-iyL5/">Armand Ruiz</a>, VP of AI Platform at IBM, puts it: <em>“Enterprise IT is not ready; most data isn’t even accessible, let alone AI-ready.”</em> <strong>Poor quality, incomplete, or biased data</strong> can undermine automation efforts, leading to inaccurate outputs and eroded trust in AI decisions. After all, <em>“an agent is only as trustworthy as the data it’s trained on.”</em><br>Beyond data readiness, there’s a significant <strong>skills gap</strong>. Most DevOps professionals don’t yet have expertise in machine learning, prompt engineering, or AI model management (model inferencing, quantization etc.). To use GenAI effectively, teams need training — not just to learn the tools, but to understand how to write meaningful prompts, interpret AI-generated insights, and integrate them into existing workflows.<br>Integration itself poses another challenge. <strong>Incorporating GenAI tools into legacy systems</strong> and mature CI/CD pipelines requires careful planning to avoid compatibility issues and disruptions. Finally, with AI agents often handling sensitive system data, privacy and security must be treated as first-class citizens, with robust compliance, governance, and security controls in place from the start.</p><p>These challenges can’t be ignored — they demand leadership. As an AI leader in the DevOps space, your role is to step up in new ways:<br><strong>Upskill relentlessly.</strong> Understand how AI works and what powers it. Your team doesn’t need to be data scientists, but AI fluency is no longer optional.<br><strong>Own your data.</strong> AI is only as smart as the signals it receives. Clean up your logs, tag incidents with intention, and feed it truth — or risk watching it learn the wrong lessons.<br><strong>Define success clearly.</strong> Your definition of success will set the direction, guide decision-making, and ensure continuous progress.</p><p>Lead with intent. Shape how AI is used in DevOps.<br>And above all, don’t forget to enjoy the journey — we’re witnessing one of the most exciting and transformative shifts in technology in decades, and you have a front-row seat.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=502a2ece61b1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Building Consensus | 5 Challenges on the Way to Success]]></title>
            <link>https://livnats.medium.com/building-consensus-5-challenges-on-the-way-to-success-ef1f3b4f7b84?source=rss-7174840c036e------2</link>
            <guid isPermaLink="false">https://medium.com/p/ef1f3b4f7b84</guid>
            <category><![CDATA[open-culture]]></category>
            <category><![CDATA[consensus]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[team-collaboration]]></category>
            <category><![CDATA[management]]></category>
            <dc:creator><![CDATA[Livnat Peer]]></dc:creator>
            <pubDate>Wed, 27 Jan 2021 13:48:15 GMT</pubDate>
            <atom:updated>2021-01-27T13:48:15.425Z</atom:updated>
            <content:encoded><![CDATA[<p>Written by <a href="https://medium.com/u/b99f290513cb">Angus Thomas</a> &amp; <a href="https://medium.com/u/7174840c036e">Livnat Peer,</a> Jan 27th 2021</p><p>Open communication is a pillar in open organization culture. Empowering people to debate ideas, to express controversial opinions and to listen to each other is core for enabling good ideas to emerge, for making more well informed, more thought through decisions.</p><p>A useful tool to have in your toolbox, in this context, is the ability to drive a conversation towards consensus. This post covers a few common traps one can fall into, while trying to build a consensus.</p><p>Before diving into the challenges in the process of building consensus, let’s describe the difference between debating an idea and building consensus. Debating an idea means that various people have an open conversation in which they potentially express different opinions, either lending their support, or expressing their reservations.. However there is no requirement that most people end up agreeing to a single view/solution by the end of the discussion. Achieving consensus, on the other hand, requires that there is some sort of agreement at the end of the discussion.</p><p>We’d like to acknowledge that building consensus is not always needed. In organizations, including in an open organization, normally there is an authority which is responsible for making the decision at the end of the day. This authority is the one to decide when an idea is debated for long enough, when the potential of new ideas to emerge is exhausted or not needed any longer and, most importantly, when it’s time to make a decision and move forward. So actually, often, all is needed, is an open discussion with the authoritative figure making a decision at the end and not necessarily achieving consensus.</p><p>Even though achieving consensus is not required in some situations, it doesn’t mean it’s not the right thing to aspire for. The process of building consensus can deliver several benefits. It can lead to more in-depth discussions, more active listening, and can help people to understand better the strengths and weaknesses of those ideas. Working to build consensus fosters greater transparency and helps in nurturing healthy communities. Reaching consensus means that more people are on board with an idea, which can later translate into more people spreading the idea, advocating for it and working to make it successful. Reaching consensus enables you to move forward, it means you dissolved the critical objections and you can now focus on implementation or making the next set of decisions.</p><p>With all the benefits that a consensus-based process can deliver, we’re sure you are going to try and achieve consensus in the next meeting you’re on. Before you do it just take a few minutes to read through the challenges and what can go wrong when striving for consensus -</p><ol><li><strong>Miscommunication of goals </strong>-</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/389/1*vfqRb2c4PQnVoyEU1-WH0Q.png" /></figure><p>This one is about setting expectations ahead of time.</p><p>In some cases one is interested in hearing different opinions and not necessarily to get to an agreement on what is right/wrong nor to decide what to do. In such a case explaining that you are not seeking to get everyone to agree (not seeking to build consensus), that at the end of the process you are likely to take the data and do something with it (would be good to explain what) can take you a long way. Making it clear you don’t expect a decision to come out of this discussion can prevent frustration and disappointment. People are likely to understand such a goal and to be less frustrated if you end up operating ‘against’ their recommendation or opinions. Explaining you are not seeking to build consensus can make the whole process much better to all participants. Including you.</p><p><strong>2. Mandating a consensus</strong> -</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/112/1*e8o22Y1PHJpyD3oPSmeK2w.png" /></figure><p>One should be careful of using authority, consciously or unconsciously, when building consensus. When a figure with authority is driving a discussion and is interested in getting a broad agreement on a topic but people in the room don’t seem to agree, either amongst themselves, or with the particular view advocated by the authority figure, there is a trap to be wary of. If the figure keeps pushing on the idea, in some cases, an apparent consensus could be reached but it would be very much aligned with what the authoritative figure had in mind.</p><p>In such cases most people will just stay quiet and accept the opinion of the authority figure. This is about mandating a consensus and not about genuinely building one. The consensus reached in this case is ‘fake’ and doesn’t really convey the benefits that come with genuine consensus. People will keep questioning the decision, mostly not publicly but in private conversations, their motivation to spread the idea will likely be questionable and they’ll be less likely to bring enthusiasm into the process. There are various ways to avoid this trap. First, one has to be aware of her/his authority and how it can influence a discussion. The more authority a figure has the more careful/aware they should be when promoting an idea in a consensus building discussion. An alternative to avoid this trap is to be open about the situation and reinforce the openness of the process, making people feel empowered to express disagreement in the conversation. Another alternative is to find champions and advocates for the idea within the team, ahead of time, and encourage them to speak up. People are likely to feel more comfortable responding frankly to their proposals than to proposals coming from an authority figure. It’s not perfect, but this can be an effective way to see how people really feel about an idea when its not coming from an authoritative figure. There are many ways to bypass this trap however the starting point is to be aware of the situation and the power balance in the room.</p><p><strong>3. Being right </strong>-</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/212/1*VlGaDJ5_GG7gdw_cUe38Xg.png" /></figure><p>Some people confuse consensus with being right. The fact that one was able to build a consensus around an idea doesn’t guarantee that the idea is the right thing to do or that it is risk free. Obviously having an open conversation about an idea makes the chances of getting it completely wrong less likely, however less likely doesn’t mean it can’t happen. If the idea is risky or in the domain of exploring new territory there is a good chance you are getting it (partially?) wrong. Building a consensus can give an illusion of confidence in the path taken and this is a trap to be aware of. The way around it is simple, not to rely on consensus alone too much nor confuse it with being right. It’s wise to keep asking questions along the way and to see if a change or adjustment are required.</p><p><strong>4. Reducing responsibility </strong>-</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/121/1*O2sJeg3jijJWDULzqc4P6A.png" /></figure><p>This one is simply a flawed motivation for building consensus. The trap is to think that by building a consensus and having a broad agreement on an idea, leaders can avoid responsibility if things don’t go to plan. It should be stated explicitly — by reaching a consensus the responsibility is not removed from the owner of the decision. At the end of the day there is a person who is responsible for the strategy or the decision at hand, and even when reaching a consensus that person would be held accountable to the results of the path taken. So build consensus to reach a better path forward rather than to escape responsibility for the path taken.</p><p><strong>5. Moving forward</strong> -</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/179/1*hytZu8jGyfKBGz-mBSoESg.png" /></figure><p>Sometimes a consensus cannot be reached, there are honest differences of opinion within the group and listening respectfully to each other’s view’s doesn’t lead to either side changing their mind. That’s not a failure, and shouldn’t be seen as a problem. We can’t require people to have opinions other than the ones that they do, and we should be very wary of asking people to act as if their opinions aren’t what they truly are. In those cases, where no consensus emerges, it is up to the leader to decide. These are the times when the need for a leader is made clear. Often, rather than feeling powerful at those times, we can feel uncomfortable. Working with the opinions of our teams, moving forward with their consensus positions, gives us the reassurance of the agreement of others. When that agreement doesn’t exist, we can’t let our teams become paralysed. We need to decide, so that the work of the team can move on. Dealing with the disagreement, leading the team forward based on our own best judgement, and taking responsibility for the team’s direction is a vital part of the role of a leader.</p><p>Building consensus is an art. Once achieved it can lead to great results. However, when using it for the wrong reasons, or in the wrong way, it can be a destructive process.</p><p>Build consensus and be smart about it!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ef1f3b4f7b84" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Management Nuggets]]></title>
            <link>https://livnats.medium.com/management-nuggets-5f7943d8ebe8?source=rss-7174840c036e------2</link>
            <guid isPermaLink="false">https://medium.com/p/5f7943d8ebe8</guid>
            <category><![CDATA[software]]></category>
            <category><![CDATA[management-and-leadership]]></category>
            <category><![CDATA[managment]]></category>
            <category><![CDATA[bottleneck]]></category>
            <category><![CDATA[engineering]]></category>
            <dc:creator><![CDATA[Livnat Peer]]></dc:creator>
            <pubDate>Mon, 08 Jun 2020 09:14:37 GMT</pubDate>
            <atom:updated>2020-06-08T09:14:37.257Z</atom:updated>
            <content:encoded><![CDATA[<h4><strong><em>Bottlenecks and Liaisons to the rescue</em></strong></h4><p>A week ago I had a chat with a friend of mine about a challenge he is facing in his RnD organization. We didn’t discuss anything new or surprising, actually I’d say that if you are a manager it’s very likely that you saw this or something similar before.</p><p>We were discussing a situation where, within the organization, one team provides a service to other teams and becomes a bottleneck.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/685/0*HRpjzWk0qoi1Fmfc" /></figure><p>For example Let’s say we have a CI-Infra team. CI-Infra is building CI libraries that are used by other teams for coding advanced testing scenarios. The CI-Infra team has responsibility to maintain these libraries, to satisfy the requirement for new capabilities that are coming from the various teams using their libraries and to innovate or keep up with the CI trends in the market.</p><p>Very soon the teams that consume services from the CI-Infra team are not happy, either because there are bugs in the packages that are not being fixed fast enough, or because the new capabilities they asked for are taking too long to develop, or for any other reason that gets them blocked on the CI-Infra services.</p><p>A common solution is to add more resources to the CI-Infra team, so they can meet the various expectations from the team and handle the growing workload.</p><p>I’d like to offer another possible solution to the problem described above. We call it the ‘liaisons model’.</p><p>The proposal is to add people to the teams that consume the services from CI-Infra. This should be done instead-of or in-addition-to adding more resources to the CI-Infra team itself. The embedded people in the other teams are the CI-Infra liaisons. The liaisons are members of the team that hires them (their ‘home’ team) and are not part of the CI-Infra team. However the liasons are working on CI-Infra tasks. More specifically, their tasks are being prioritized by their home team and are on the CI-Infra backlog.</p><p>The liaisons’ main job is to resolve the bottleneck the home team experiences with the CI-Infra team. They would actively gain expertise, contribute code and participate in the CI-Infra work. The key difference from other CI-Infra members is that they are there to serve their home team. They get priorities from their home team and work on bug fixes or new features that are blocking their home team.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*o4dbkbmnPZIEmvyN" /></figure><p>On top of resolving the bottleneck, the liaison approach gives you resource flexibility. You can identify when CI-Infra is no longer a bottleneck for the home team and maybe there are no additional requirements for services from CI-Infra, or at least not enough to justify a full time liaison. Then you can assign the liaison to work on additional tasks that are serving your home team needs (not necessarily related to CI-Infra) or if it’s better for your organization to move the liaison to be CI-Infra liaison of another team that needs it more. The flexibility is built into the model, It just needs to be used wisely.</p><p>A key aspect for the liaison model’s success is endorsement by the CI-Infra team. The CI-Infra team is expected to mentor and help the liaisons, which requires non-trivial investment. The liaisons are not CI-Infra team members which makes the CI-Infra engagement to this model more challenging. The good news is that there are some things you, as a manager, can do to help make this successful.</p><p>There are multiple ways to help support and cultivate good relationships between the liaisons and the CI-Infra team.</p><p>It is recommended to make the liaisons’ nomination a mutual decision of the home team and the CI-Infra team. If the home team is hiring a new person to be the liaison they should get the CI-infra team to take an active part in the screening process. If they are assigning someone from the existing team they should talk to the CI-Infra team and see if there is a personality match there. The relationship of the liaisons with CI-Infra is key for the success of this approach.</p><p>There is an option to make a member of the CI-Infra team a liaison, by moving him/her to the home team as liaison and hire someone new into the CI-Infra team. Using this approach leverages the trust and maybe friendly relationship the person has with the other CI-Infra members. Obviously, this approach doesn’t scale and can only be used a very limited number of times, but it can be a good start when adopting the liaisons model.</p><p>Another approach would be to assign a senior mentor from within the CI-Infra team to the liaison. Associate the liaison’s success with the mentor’s success, that should motivate the mentor to work closely with the liasson. Also if you choose a senior enough team member to be the mentor, there is a chance the other CI-Infra team members would follow the mentor lead and the endorsement of the liaisons model overall would be more smooth.</p><p>Allocating some capacity of the liaisons’ time, to contribute back to CI-Infra, can be valuable to the motivation of the CI-Infra team. Most of the liaison’s time is spent on working on items for their home team but there can be an additional commitment to work on the CI-Infra backlog and help them make progress on their goals.</p><p>As you see there are various ways to help build good relationships between the CI-Infra team and the liaisons and I’m sure there are many more things you can do. If you choose to apply this model, the important point is to consciously spend time thinking on these relationships to make the change successful.</p><p>Another piece of advice that is applicable to this model as well as to many other models out there: Make sure you have a clear definition of the roles and the expectations from each party. Define clearly what the liaison’s role is in your organization, along with the role of CI-Infra team members, where they should collaborate and what expectations there are from both parties. The more clear the definitions are the less likely people would step on each other’s toes.</p><p>The liaison model doesn’t fit all teams, if there is a small team with only a handful of requirements from the CI-Infra team, maybe adding a liaison to that team is not the most suited solution. Remember the liaisons come with a toll on the CI-infra team, because the training burden is on them, so it might not make sense to ask them to mentor a person just to serve a few, far apart, requests. In some cases it makes more sense to ‘wait’ in the queue for the CI-infra team to handle requests.</p><p>I used the liaisons model in the past and saw it work. Obviously it is not a good fit for all cases. In some cases it is better to scale up your CI-Infra team, in others to use the liaisons model a third option is to combine the two. Choose wisely what can work for your organization. Whatever you choose, get people around you to be engaged with the success of the process.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5f7943d8ebe8" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>