<?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 Gianluca Brugnoli on Medium]]></title>
        <description><![CDATA[Stories by Gianluca Brugnoli on Medium]]></description>
        <link>https://medium.com/@lowresolution?source=rss-693ebd7d07c2------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*XA5Kw7ut0iwgsjVjRCT_GA.jpeg</url>
            <title>Stories by Gianluca Brugnoli on Medium</title>
            <link>https://medium.com/@lowresolution?source=rss-693ebd7d07c2------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 12:03:33 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@lowresolution/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[The checkpoints of the AI User Experience]]></title>
            <link>https://medium.com/@lowresolution/the-checkpoints-of-the-ai-user-experience-421c7ecfcf5a?source=rss-693ebd7d07c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/421c7ecfcf5a</guid>
            <category><![CDATA[service-design]]></category>
            <category><![CDATA[ux-design]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[ux]]></category>
            <dc:creator><![CDATA[Gianluca Brugnoli]]></dc:creator>
            <pubDate>Wed, 15 Apr 2026 09:50:00 GMT</pubDate>
            <atom:updated>2026-04-15T09:51:56.023Z</atom:updated>
            <cc:license>https://creativecommons.org/licenses/by-nc-sa/4.0/</cc:license>
            <content:encoded><![CDATA[<p>Improving human–AI collaboration while preserving user control</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*C1uC6KR4CiU7LykV" /><figcaption>Photo by <a href="https://unsplash.com/@project2204?utm_source=medium&amp;utm_medium=referral">Leif Christoph Gottwald</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>AI is changing the way humans interact with digital systems. So far we have been designing software tools for people, instruments for doing things. We are now entering an era where AI services serve us, doing things autonomously on our behalf.</p><p>This essay explores what this means for designers, for the field of user experience, and for the users involved in these new digital experiences.</p><p>It’s also a reflection: this change is not an incremental evolution of what we already do. It’s a break with the past that compels us to re-examine some of the assumptions our industry has accumulated over the last thirty years.</p><h3><strong>How AI is changing the role of the user</strong></h3><p>Let me offer a scenario that is, as of today, entirely plausible.</p><p>You need to travel from Milan to Amsterdam for a two-day meeting next Thursday. In the recent past, you would have opened a tab, navigated to a travel booking website, entered your origin and destination in a webform, selected the date, filtered by price, compared options, chosen a flight, selected a seat, entered your personal details, paid, then repeated the process for a hotel, then again for a transfer. At each step, the software guided you through a predefined process and offered you a defined set of options. You evaluated. You clicked. You moved forward. You were, in the fullest sense, the software operator.</p><p>Today, you can open an AI chat interface and type: “<em>Book me a flight to Amsterdam next Wednesday, back Thursday evening, window seat, I prefer early morning departures, and I usually stay near the city centre.</em>” And the agent, having access to your calendar, your company payment details, your travel preferences, and a real-time connection to booking systems, handles the request. The final itinerary arrives. The confirmation is in your inbox.</p><p>You did not navigate a website. You did not compare a list of options. You did not filter. You expressed an intention, and something else, interpreted it, elaborated it, and resolved it providing you with a solution.</p><p>This is a fundamental shift in the digital user experience. For a long time, our design goal has been to create software interfaces operated by users. The human is the agent; the software is the instrument. AI now inverts that model: <strong>the AI is the agent, the human is the requester</strong>.</p><h4><strong>From self-service to being served</strong></h4><p>In the digital world users interact with online services in <strong>self-service mode</strong>. They serve themselves, autonomously, searching for information, making choices and taking actions using the digital tools at their disposal.</p><p>This is normal today, but before it was completely different. Digital applications have replaced a range of traditional human-mediated services. Customers of previous generations were accustomed to being assisted by a real-life representative, whether at a bank, a travel agency, or a downtown store. When they needed to book a flight, for example, they didn’t have to do it by themselves. Instead, they received assistance and advice directly from an industry professional who handled (almost) every step of the process for them, delivering a complete solution.</p><p>Today, the customer has become a digital user who achieves the same outcomes by interacting with digital platforms in a self-service manner, without the help or assistance of a sales associate or support representative. The self-service model of digital services is made possible by the <strong>fundamental role played by user experience design</strong>: if the user is unable to navigate an app and complete tasks independently and easily, the product fails. The quality of the user experience (and the user interface) is a critical business success factor in a world where users must interact with digital services on their own, without assistance.</p><p>In the world of AI agents, the role of the “user” shifts: it no longer refers to an individual operating software, through structured interfaces following a sequence of tasks. Users no longer interact with an application in self-service mode. <strong>They have returned to being customers with needs and goals who are being served</strong>. In this case, no longer by a human operator, but by an artificial intelligence agent. The user sets the goal, the agent finds the path and provides a solution autonomously.</p><blockquote>From the user experience perspective, AI brings users back into the world of assisted services, where someone, or something, performs tasks on their behalf.</blockquote><p>The AI agent, in its social logic, is very different from software as we have conceived it until now. It is much more like a personal assistant, a concierge, or a qualified professional acting on the user’s behalf. This is not merely a metaphor: it reflects a real shift in how agency is distributed in digital systems.</p><p>As AI becomes more autonomous, humans must not be reduced to passive users of automated outcomes. The key challenge is <strong>preserving human agency, involvement and control</strong>. This requires redefining the user’s role as an active participant in AI experiences, fundamentally reshaping what we design, and reframing who those systems are ultimately built for.</p><h3><strong>AI is changing how we interact with digital services</strong></h3><p>A piece of software is a system with a predefined logic. Every action it can perform is enumerated in its codebase. Every user case and workflow is designed in advance. The user interface is, essentially, a map of the software’s capabilities, a visual representation of what is possible, that guides user interaction in a form that humans can understand and navigate without knowing how the underlying system works. When you book a flight on a legacy booking platform, you are moving through a decision tree and a set of user flows that was designed by a team of product managers and designers years before you arrived. The steps are fixed. The options are finite. The outcomes are predictable and consistent.</p><p>This determinism is not a flaw. It is the design. The entire discipline of UX design, as it has evolved over the last two decades, is built on top of this determinism.</p><p>An AI agent has a different operating model. Its behaviour is not the output of a deterministic decision tree, but of a <strong>probabilistic model that interprets natural language</strong>, draws inferences, selects among available tools and data, and generates responses that are shaped by context, prior conversation, and the specific phrasing of the user’s request. Two users asking for essentially the same thing to an online service may receive different responses. The same user asking the same question twice may get different answers. An AI system adapts its communication style and its decision-making process to the situation and the data available in that moment.</p><p>This is both a potential and a design challenge. The potential is clear: an AI agent is <strong>highly flexible and can handle an almost unlimited range of requests</strong>, adapting to edge cases that no designer could have anticipated, and delivering results that feel personalised, rather than simply being selected from a limited menu of options. Any workflow and sequence of tasks is <strong>radically simplified</strong>: users no longer have to go through lengthy online processes, dealing with multiple interactions and complex interfaces.</p><p>The challenge is also obvious: how do you design an experience for a system that, by its nature, does not behave the same way twice? How do you support feature and content discovery, giving users a sense of what is possible when there is no interface to explore, no menu to browse, no feature list to consult? How do you help users understand what went wrong, or what went right, when the logic behind any given output is not a traceable decision tree but the emergent output of a large language model reasoning across hundreds of variables?</p><h3><strong>AI user experience: three key challenges</strong></h3><p>With AI, the challenge is not to make systems easier to use, but to make them <strong>reliable and trustworthy</strong>. Once we accept that AI services are based on autonomous agents and that users must retain an active role, a different class of UX problems comes into focus, problems no algorithm alone can solve.</p><h4><strong>Control: when there is no undo</strong></h4><p>Usability research on AI-powered interfaces shows a consistent behaviour: when an AI produces a wrong or unexpected outcome, users rarely try to fix it incrementally continuing the chat conversation. It’s too complicated. Consumers reset the conversation and start over with a new request. This is a rational response to systems whose state is opaque and hard to reason about. Unlike traditional software, where changes are visible and reversible, conversational AI distributes state across a long dialogue history. Correcting an error requires understanding not only what went wrong, but also why. This cognitive burden breaks the promise of a simple and natural experience.</p><p>As AI agents gain authority to take real-world actions, this becomes a serious design risk. The solution is architectural, not technical: we need <strong>decision points</strong> that must be explicit, surfaced before commitment, and paired with clear mechanisms for review, modification, and cancellation.</p><blockquote>Autonomous action is a feature; autonomous action without human supervision is a liability.</blockquote><h4><strong>Transparency: inside the black box</strong></h4><p>If an AI makes a decision on your behalf and you cannot understand why, trust remains partial at best. This is not resistance to AI, but a reasonable response to being asked to accept outcomes without access to their reasoning. Transparency and explainability are not optional; they’re a <strong>foundational component of the AI experience</strong>.</p><p>Effective transparency is not about exposing models or confidence scores, but about explanations in human terms. People don’t want to be passive AI users, they need to understand which factors mattered, how they were weighed, and how the outcome aligns with their goals. They also need to know the sequence of human and agentic decisions that led to the outcome, when necessary.</p><p>Waymo, for example, displays its sensor perception alongside the road map on passenger screens, visualising in real time the the world around the car with the route. By exposing how the vehicle detects and interprets the surroundings, it turns the “AI black box” into a legible and predictable experience, building trust through transparency.</p><p>Research consistently shows that when AI reasoning is communicated clearly, users report higher confidence, agency, and willingness to engage. Transparency operates at three levels: decision transparency (why this outcome), process transparency (what the agent is doing now), and accountability transparency (how a decision can be reconstructed after the fact). Together, these ensure that users retain authority over actions taken in their name.</p><h4><strong>Consistency: the agent as a reliable partner</strong></h4><p>Traditional software earns trust through consistency. AI agents, by contrast, are probabilistic, and behavioural variability can feel unsettling to users. The challenge is not to eliminate this variability, but to constrain it within a <strong>coherent and recognisable pattern of behaviour</strong>.</p><p>In agentic UX, consistency is achieved through a stable persona, predictable interaction patterns with clear confirmation and validation moments, and visible memory of user preferences across sessions. AI systems do not follow the same process every time, and the data shaping each response may change, but they can still offer consistent control points where users can check status, validate progress, provide feedback, and intervene when needed. These intentional moments of interaction improve transparency and user agency, making the overall experience feel more predictable, manageable, and trustworthy.</p><h3><strong>The new AI Journey: from frictionless flows to checkpoints</strong></h3><p>For the past decade, the prevalent goal of any experience designer has been the <strong>frictionless journey</strong>. Remove every obstacle between the user’s intention and its fulfilment. Eliminate unnecessary steps. Pre-fill forms. Reduce the number of screens and clicks. Make the path from intent to outcome as short and as smooth as possible.</p><p>This aspiration is not wrong. In the context of self-service software experience, friction is the enemy. Every unnecessary click or form field is a potential drop-off point. Every additional step is a threat to conversion.</p><p>In the agentic experience, the fastest path is no longer the best path. In this scenario users are no longer in control of the process and sometimes not even part of it. They are being served and the agent handles it in its own way. They have no agency. Unless we deliberately design points of human re-entry into the process, the user can go from intention to outcome without ever being consulted.</p><p>This reframes the customer journey design problem entirely. We are no longer asking: “where is the friction, and how do we eliminate it?” We are asking: “where are the moments of human decision, and how do we preserve them? How can we keep the user in control?”</p><p>The new customer journey map and service blueprint no longer aim to minimize interactions and touchpoints. Instead, they focus on identifying and designing appropriate <strong>UX checkpoints </strong>of the AI service experience where collaboration between humans and AI occur.</p><p>These checkpoints are specific moments in the journey where the AI flow must pause to accommodate human judgment, address high-stakes decisions, or obtain required verification. Rather than removing steps, we <strong>intentionally add new ones</strong>: validation moments designed to improve the user experience, keep humans in control, and provide transparent mechanisms for shaping system outcomes.</p><blockquote>Intentional interactions at critical junctures are now the core of the AI experience rather than barriers to eliminate.</blockquote><p>Well-designed UX checkpoints in AI service experiences are meant to address the three key UX challenges previously outlined: keeping the user in <strong>control</strong>, ensuring <strong>transparency</strong>, and maintaining <strong>consistent</strong> interaction patterns.</p><p>Good examples of UX checkpoints include moments when the chatbot asks a clarifying question, provides intermediate results and offers options for the next action, explains the reasoning behind a suggestion and a result, visualise the state of the system, or prompts the user for feedback.</p><p>UX checkpoints are the steps in which the user can have an active role, express preferences, supervise the process to the outcome, corrects and verifies the choices made and builds (or fails to build) trust in the system acting on their behalf.</p><h3><strong>UX strategy: checkpoints for a more effective human-AI collaboration</strong></h3><p>AI-based digital services rely on a hybrid collaboration between humans and machines, often formalised through a human-in-the-loop approach. UX checkpoints are the critical moments where this <strong>collaboration is enacted</strong>.</p><p>A checkpoint-based approach shifts UX in AI systems from optimising flows for efficiency to deliberately orchestrating moments of control, evaluation, and trust in which the user plays a new active role. Strategically, this is a significant change in how autonomous services are conceived, designed and measured.</p><p>From a strategic perspective, checkpoints also redefine how <strong>value is created</strong> in AI services. <strong>Effective collaboration </strong>becomes the central strategic goal: the quality of the experience is no longer measured only by speed or friction reduction, but by how effectively the system <strong>collaborates with the human user to generate business and customer value</strong>.</p><p>In this way, human oversight is explicitly integated into the AI service model. This has direct implications for risk management, especially in high-stakes domains such as finance, healthcare, legal workflows, autonomous driving or any context where AI decisions carry meaningful consequences. Over time, this can become a differentiator not just in usability, but in perceived responsibility, transparency, and long-term user adoption.</p><blockquote><em>UX checkpoints are the strategic foundation of human-AI collaboration. They transform AI systems from optimized flows into governed interactions, where control, evaluation, and trust are intentionally designed into the experience.</em></blockquote><p>UX checkpoints act as building blocks for defining the collaboration model of an AI service, which must be designed on a case-by-case, industry-by-industry basis. In consumer contexts, well-designed checkpoints are not only verification steps but also moments of engagement and reassurance for users. In business contexts, checkpoints serve an additional validation step where human expertise, judgment, and domain-specific context are required to supervise the process and guide outcomes before the system proceeds.</p><p>For an effective human-AI collaboration, UX checkpoints should never feel like interruptions or obstacles. If they are clumsy, ambiguous, or poorly timed, they risk becoming sources of friction or frustration, moments where the user hesitates, loses confidence, or disengages. Instead, they should feel like <strong>natural, even welcome, parts of the flow</strong>: opportunities for the user to steer, to reflect, to confirm or correct and explore alternatives.</p><p>Designing these supervision moments requires clarity in both the interface and the intent. In a UX checkpoint <strong>the visual user interface can be relevant again</strong>, playing a role in presenting options, actions, and messages to the user in a simple and clear way. The user should be engaged at the right moment and with the right interaction. The user interface should enable quick and confident action, without overwhelming the user or demanding unnecessary effort and complicated inputs.</p><blockquote>Ultimately, a good UX checkpoint is not just a UI element, it’s an opportunity to create a delightful and engaging customer experience and a moment of trust-building.</blockquote><p>These steps are also key moments in which the AI experience can be <strong>personalised and adapted in real time</strong> to the user context.</p><h3><strong>Designing UX checkpoints</strong></h3><p>Identifying and designing an effective checkpoint is clearly a UX task: it must feel helpful rather than intrusive, transforming a potential point of friction into a moment of empowerment that builds trust and reliability.</p><p>UX checkpoints can be identified and mapped along the journey and designed using <strong>established user-centred practices</strong>. We still have to understand users, their goals, needs, and expectations, as well as the type of collaboration they want with AI in different use cases. This understanding is then translated into UX and UI solutions that keep users in control and provide transparent mechanisms for shaping system outcomes.</p><p>A few principles for designing effective UX checkpoints in AI-powered CX:</p><ol><li><strong>Make checkpoints deliberate and contextual</strong>: Don’t hide or skimp on interaction moments. Identify points in the journey where a human decision or input is vital, be it due to high stakes, potential errors, or user comfort, and mark those as key checkpoints. For each checkpoint, design a clear and concise user interaction (a question, choice, or confirmation) that fits naturally into the flow. Examples: Klarna’s bot explicitly hands off when it senses trouble, and Duolingo offers an “Explain My Answer” button after each question. These aren’t ad-hoc interruptions; they’re planned moments that serve a purpose.</li><li><strong>Adapt the checkpoints to different context and user needs</strong>: Not all users want the same level of oversight in their interactions with AI agents. Some may prefer a streamlined, low-friction journey with minimal interruptions, while others may feel more comfortable with multiple checkpoints that offer opportunities to review, confirm, or adjust the AI’s actions. This variation can depend on the user’s familiarity with the system, their risk tolerance, or the nature of the task at hand. Context also plays a critical role. In B2C environments, such as e-commerce or travel, users may expect speed and convenience, with checkpoints appearing only at key decision moments (e.g., payment confirmation). In contrast, B2B scenarios often involve higher stakes, longer decision cycles, and a greater need for traceability and control. More frequent and detailed checkpoints may be not only expected but required for compliance or internal governance. Users should be able to set their own preferences, where possible, or customise the interaction model based on user type, task complexity, or industry norms. This ensures that the AI experience feels appropriately responsive, neither overly intrusive nor dangerously opaque.</li><li><strong>Maintain consistency in AI behaviour around checkpoints</strong>: Consistency is key to user trust. Define clear policies and guidelines for when each checkpoint is invoked, and stick to them. For instance, if your travel-booking AI always asks the user to confirm purchases above a certain amount, don’t skip that step on the 10th or 50th booking. Consistency also means the AI should explain or confirm in a similar manner each time, using a recurring user interface pattern with familiar language and tone. As a study on AI trust shows, predictable behaviour, even if slightly less optimal, is preferable to users over random high variance. By keeping checkpoint interactions consistent, you allow users to develop an accurate mental model of the AI’s operation, which is essential for trust.</li><li><strong>Leverage human oversight as a learning opportunity</strong>: Whenever a human intervenes, use it as a chance to improve the AI. Capture why the checkpoint was triggered and what the human did, then refine your system. Over time, this reduces the need for intervention on the same issues and makes the UX checkpoints behaviour smarter and less intrusive.</li></ol><h3><strong>What this means for designers</strong></h3><p>For the last two decades, the primary focus of user experience design has been on interface design and the structuring of user flows. These skills are still valuable, and they will not disappear, because even in an agentic world, visual interfaces do not vanish. Humans do need visual feedback about the state of the system. The user interface is not obsolete, it is repurposed. But it’s no longer the primary design focus.</p><p>In the agentic era, the primary design goal should be building a <strong>positive collaboration</strong> between a human and an AI autonomous system. And shaping that relationship with UX checkpoints draws on a much broader set of disciplines beyond screen design.</p><p>It draws on <strong>service design</strong>, the discipline that has always been concerned with the end-to-end experience of a service, across channels and touchpoints. Service designers know that the experience of a service and value creation is not reducible to its individual touchpoints: it is constituted by the quality of the interactions at those touchpoints, the coherence of the journey between them, and the trust that accumulates or erodes across the full arc of the relationship. Moreover, service designers have a robust systemic thinking, that maps directly onto the challenge of designing effective human-AI collaborative experiences and strategies.</p><p>It draws on <strong>ethics and governance</strong>, the capacity to reason carefully about where autonomous action is appropriate and where it is not, about what information users deserve to have access to, about the conditions under which a system should refuse to act, ask for confirmation, and when and how to involve a human in the loop. These are not engineering decisions. These are UX design decisions, and they carry moral weight.</p><p>And it draws on something that might be called <strong>trust architecture</strong>, the deliberate design of the mechanisms through which a user’s confidence in an autonomous system is earned, maintained, and, when necessary, repaired. Trust is not a feeling that emerges spontaneously from good technology. It’s built through a good user experience: consistent behaviour, legible reasoning, and the felt sense that the system is operating in the user’s interest rather than against it. In the agentic world, designing for trust is as fundamental a design objective as designing for usability was in the self-service world.</p><p>Finally, there is an overarching question that matters more than any specific skill, tool, or product: how the UX mission itself must evolve.</p><blockquote>Preserving human agency in the age of artificial intelligence is the new UX imperative.</blockquote><p>Designing AI experiences that are transparent, honest, and trustworthy is therefore not only a UX responsibility, but a fundamentally ethical challenge.</p><p>In the digital self‑service era, UX designers are, at their best, advocates for users inside product organisations. By reducing interaction complexity and improving ease of use, they aim to make technology products more accessible, more inclusive, and more democratic.</p><p>In the agentic era, that advocacy becomes both more urgent and more complex. As digital products turn into autonomous services, designers must actively defend the presence of meaningful UX checkpoints, deliberate moments of interruption that keep the human in the loop, enabling control, transparency, and judgement.</p><p>From the UX point of view these checkpoints are what make AI systems comprehensible, accountable, and reliable. From the strategic point view they are the way in which we preserve human agency. Usability and accessibility are no longer achieved by removing interface complexity or streamlining workflows, but by adding intentional, clear and useful UX checkpoints that keep users in control of the interaction with the AI agent.</p><p>Seen in this light, this is not a break from our discipline, but a more ambitious continuation of it.</p><p>UX is what turns technology into products people can meaningfully use. The responsibility of designing technology that serves humans, rather than bypassing them, has never been greater.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=421c7ecfcf5a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Designing smart experiences]]></title>
            <link>https://medium.com/@lowresolution/designing-smart-experiences-a6e675b414ec?source=rss-693ebd7d07c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/a6e675b414ec</guid>
            <category><![CDATA[internet-of-things]]></category>
            <category><![CDATA[service-design]]></category>
            <category><![CDATA[ux]]></category>
            <dc:creator><![CDATA[Gianluca Brugnoli]]></dc:creator>
            <pubDate>Tue, 17 Feb 2015 08:20:25 GMT</pubDate>
            <atom:updated>2018-09-01T17:33:32.960Z</atom:updated>
            <content:encoded><![CDATA[<h2>Designing Smart Experiences</h2><h4>The smart and invisible future of interactions and services.</h4><p>A new generation of products and services based on smart technologies is about to change the world around us. Connected objects, smart assistants, sensors, beacons, big data and artificial intelligence are revolutionising not only many industries and business sectors but also the way we will interact with an emerging world of smart devices and services.</p><p>This new generation of services, products, applications and tools will deliver to the user a new kind of experience: a “<em>smart user experience</em>”.</p><p>What is a “smart user experience”? Briefly, it is a new class of experiences based on smart technologies: ”<em>technologies with the ability to sense changes in their circumstances and execute measure to enhance their functionality under new circumstances.</em>” (1)</p><p>A smart user experience is enhanced by sensors, connected things, AI algorithms and any other technology, that is able to listen, learn and adapt automatically in real time following the users behaviour, their context and any other specific condition and rule, that influence the service delivery and outcome and, eventually, the experience itself.</p><h3>When products and services get digital, get smart.</h3><p>As soon as a product or a service combines digital technologies, sensors and network connection capabilities, is ready to get “smart”. As Michael Porter writes, this is the “<em>third wave of the IT-Driven competition</em>”, when IT and connectivity becomes integral part of the product itself (2), expanding opportunities for new functionality and capabilities that cut across and transcend traditional product boundaries. The ongoing rapid decline in the cost of sensors, combined with the similar decline of the cost of computing power, opens up many opportunities for new consumer applications of the Internet of things, changing drastically many product experiences. Almost any object could embed and integrate various sensors and get connected.</p><p>The more this technology is embedded everywhere in consumer products, mobile devices and common things, the more it becomes invisibile and deeply intertwined with our everyday experience, breaking any barrier between digital and physical domains. In the words of Eric Schmidt (3):</p><blockquote>“The internet will disappear to become part of our presence all the time.”</blockquote><p>Actually Schmidt recalls Mark Weiser (4): as technology matures, getting pervasive and connecting everything, we will experience an highly personalised and interactive world.</p><p>Converting the massive volume of data collected by sensors and mobile devices into useful results and applications is still a challenge. This is dealt with in recent developments in artificial intelligence and remote computing. Rendered by data and sophisticated software algorithms, many working “Narrow Artificial Intelligence” (also called “Weak AI”, a very specialised set of instructions focused on specific tasks) are already part of a wide array of consumer and industrial services and applications, enabling a new generation of tailored and improved services and experiences.</p><p>We are just at the beginning, most of the first commercial applications are still limited and partial, but the list of services and products enhanced by some sort of software intelligence is long and is growing every day: from Nest to Google Now and Waze. Smart healthcare systems that track and assist patients in managing their therapy and new testing systems that help doctors to decide whether to perform surgery on patients. Automatic personal financial advisors, “robo-advisors” that manage investment portfolios with a little human supervision. Agriculture tools that analyse soil hydration and environmental factors, revealing trouble spots or opportunities for better crop management. Self-driving cars, connected vehicles, and smart energy management systems. Besides many mobile smart apps and personal assistants that sort and prioritise automatically email messages, appointments and suggest what you might want to buy or the movie you may want to watch this evening.</p><p>There are massive potential innovation opportunities for many different industries and business sectors: transportation, healthcare, financial services, education, retail, entertainment, consumer products, public services and so on. A wide range of human activities and tasks will be innovated and automated by intelligent systems capable to collect, process and organise data and information better and faster.</p><h3>Next generation of services will be smart.</h3><p>The impact of the smart technologies innovation is twofold: they change the way services and products are conceived and delivered, as well as the way users interact with those services and products.</p><p>The bar of the relationship with customers is about to rise again. If the current challenge for many companies is to set up an integrated and consistent digital-physical platform for their customers across channels and touchpoints, the next challenge will be delivering a smart user experience.</p><p>To remain competitive, companies will have to rethink the customer relationship building an interactive platform able to change and adapt itself continuously, in real-time, following and engaging the customer in new ways.</p><blockquote>When everything is connected and interactive, a service can be everywhere, anytime, always ready and available for the customer, regardless of channels, physical locations and time of the day.</blockquote><p>Digital banking and ecommerce are excellent examples of this transformation and of its consequences on the service experience. Users simply don’t think about channels, locations and time of the day: they just access the service opportunistically, when they want, anywhere, with the most convenient devices available in that moment. For a service being online and accessible by a mobile device is just the first step, a natural and basic functional condition. Now users ask for a simpler, more relevant and personalised service. A smart service.</p><p>Many traditional objects and spaces, like retail and work spaces, are about to get alive and interactive, capable to detect, identify and react to the user presence, delivering contextual and personalised content and services on smartphones and other connected devices. In this hyperconnected scenario, user centrality is more important than ever: being able to follow and support people everywhere continuously and consistently will be a crucial success factor for the next generation of service experiences.</p><h3>Towards new UX challenges</h3><p>What are the key characteristics of a smart experience? Most of them are inherited from the technologies and the platforms used. Some points are not completely new for designers but get new implications in the smart services scenario. New opportunities and challenges are unfolding for user experience designers, calling into question priorities and interaction logics far from conventional screen-based interactions and graphic user interfaces. The following is an intial and brief overview and analysis.</p><h4><strong>The platform is the experience</strong>.</h4><p>The more objects are connected the more <strong>the platform and the applications available on the platform become the center of the user experience</strong>. In the “internet of everything” scenario any individual device, sensor and piece of hardware is just a resource of a distributed technical platform, often invisible (and sometimes negligible) for the user. Each connected device and every microinteraction it enables, is part of a technical ecosystem that provides the hardware platform for a software platform, which in turn enables many service solutions. Think about a connected light bulb: it is just a component of a service that is the only thing that matters for the user. The value is not delivered by the single connected object but by the software platform that combines everything (usally elsewhere, in the cloud) and that provides the user an appropriate interface to interact with the platform consistently across touchpoints and moments in time.</p><p>The ecosystem configuration can be very open and variable: some tools collect the data, some process the data and other devices provide the interface for the user. Any data collected at any point of the system becomes useful only when it is combined and processed by the software platform to deliver a specific result. The same connected gadgets and data can be used by different platforms with different purposes to deliver different solutions. The user value lies in the distributed smart service and its outcomes, not in the single smart devices.</p><h4>Physical and digital don’t matter.</h4><p>More than ever, a smart experience <strong>combines digital and physical worlds seamlessly. </strong>The underlying interactive ecosystem is an intertwined multi-device systems by definition, that combines different interaction models and inputs devices. Beyond traditional devices, the “internet of everything” includes potentially every connected device, sensor and thing (wearables, connected cars, implantable devices, accessories and so on). As a consequence, the distinction between digital services and real world touchpoints becomes less and less significant.</p><p>Ensuring the appropriate service outcome everywhere, anytime and its relevance for the context of the interaction, is critical for the relation with the user and one of the key problem for UX designers. Also a physical connected gadget and its features should be designed considering the goals and the features of the service platform it is supposed to belong, not the other way round.</p><h4><strong>Invisible interactions.</strong></h4><p>The more smart connected objects and services blend into the background of our every day life and in the world around us, the more <strong>interactions are going to be hidden, pervasive, implicit and automatic</strong>. Eventually out of the user control. In a world where sensors, interactive screens and devices are literally everywhere, most of the critical data exchange and interaction with the system that delivers the smart service is going to be less and less visible and direct. With ambient intelligence common behaviours and actions are tracked by multiple gadgets and sensors and become a source of data and triggers of many possible interactions. Crossing a doorway, for example, or walking near an object can activate various reactions, processes and feedback. Moreover the same action in different moments of the day or of the week, can activate different processes and deliver different outcomes.</p><p>The interaction with a system can be triggered by many different events, microinteractions, data and hidden controls, even without notice or a visibile feedback for the user. Multiple types of input devices can be distributed everywhere and combined together, creating thousands of interfaces of any sort, very different from the usual computer-based interaction paradigms. The triggering events of a smart experience are potentially unlimited. This is one of the key difference between traditional interaction design and designing for smart services: “<em>Without a mouse and keyboard the new input is the user’s behavior itself</em>” (5). Designers will have to rethink the human-system interaction beyond traditional screens, graphical interfaces and common input controls. Also in terms of usability, it’s not important to consider just the usability of individual UIs and connected objects but <em>interusability</em>: the distributed user experience across multiple devices (6).</p><h4>Highly personalised, nearly predictive experiences.</h4><p>A smart application delivers only <strong>what the user needs in a specific situation, following and anticipating tasks and expectations. </strong>Personalisation is not a new topic for designers and UX strategists, but in this case the experience goes beyond content and data to touch features and processes. Regardless the device or the output method, similar user actions and interactions can deliver different dynamic “smart selections” or “reactions” depending on various “smart conditions” managed by the system. Location, time of the day, user preferences, interaction history, weather and traffic condition, social connections and any other data and rules can affect the outcome of a smart service and the way it is delivered to the user.</p><h4><strong>Less, smaller and simpler interfaces</strong>.</h4><p>This point is perfectly described by Rehabstudio as “<em>more learning, less interface</em>”: the more a system can learn from the user behaviour and adapt accordingly, <strong>the less an interface is necessary or a UI can be simpler and smaller</strong> (7). Making the most of cloud services, deep learning and remote intelligent algorithms, a smart service combines data and reduces automatically the possible options, showing to the user only the content and the features that are useful or relevant for that specific context and situation (for example a route to follow, a number to call, a set of content, a notification, a set of features and so on).</p><blockquote>With smart services, live notifications, smart suggestions and dynamic recommendations are key moments of the user experience.</blockquote><p>Users do not have to search or browse an application menu to find what they need, but they simply choose within a smart selection of the best combination of content and features related to their specific situation, prepared automatically by the intelligent system. In many situations the experience will be based on live and dynamic notifications, suggestions and recommendations delivered to the user on the spot at the right time. This is the use case of smart watches, for example, and of other small devices with small screens or limited interaction capabilities (wearables, smart home connected objects and so on).</p><h4>The interaction is a conversation.</h4><p>Sometimes, <strong>the interaction with the user often is initiated automatically by the service itself</strong>, with a notification or an highly personalised screen. This is more than natural interaction, gesture or voice interfaces: it’s a new set of possibilities, logic, constraints, affordances and issues for designers. If a system or a device is given the possibility to make choices, take autonomous decisions and be active part of the interaction with the user, UX designers will have to deal with new challenges (8), both for the user interface as well as for the overall user experience strategy, content distribution and service organisation.</p><h4><strong>Remote user interfaces</strong>.</h4><p>In the everything connected scenario, <strong>user interfaces and smart devices are often decoupled</strong> and the experience is based on remote interactions, or on a combination of many types of interactions. Different devices and screens get combined and the experience is distributed: a device is used to control another and viceversa. A smart bulb is again a good example of this situation: a remote control application can manage different connected devices contemporarily often applying conditional rules. These kind of interactions can be based on proximity and near-field communication (for example, interacting with a product while standing in front of the product itself), or remote interactions (for example interacting with a system over the internet).</p><p>Many smartphone apps are the main interface for a wide array of connected objects and devices that have no UI at all, making our smartphone into a remote control for pretty much everything. Actually this is also one of the main issue of many smart platforms: if you need a smartphone (and an app) everytime you have to turn off the lights or change the temperature of the room, the experience can get cumbersome pretty quickly. As mentioned earlier, simple tasks are better managed with small and simple interfaces.</p><h4>The result is provided combining the features of many different devices.</h4><p>The user interface of smart service usually does not provide access to every single connected thing or sensor. <strong>The service result is a smart combination of the features of different devices</strong>. For example, many smart-home applications are based on “scenarios”: the user manages various scenarios (a combination of light condition, temperature and music — just to give a very basic example) and a software from an hub manages the network of connected devices to provide the desired result, dimming down the light, changing the temperature, managing music or controlling any other domestic appliances accordingly.</p><p>The “scenario” concept is pretty common in smart services: it includes conditions, rules and variables that are managed dynamically by the system itself, learning and adapting continuosly the outcomes following contextual situations and user behaviours. Scenarios (recipes, combinations and configurations) can be easily shared with other users of the system making the overall experience more engaging and social.</p><h4>Provide a smart starting point.</h4><p>In a smart service users don’t start the experience from a blank page. They <strong>start their experience from a predefined set of scenarios which evolve with the user interaction</strong>. Actually also this feature seem not very new: providing a set of predefined templates is a common solution of many digital applications and services. The difference lies in the continuous automatic adaptation of the scenarios: any predefined configuration provided by the system is not supposed to be fixed and closed, it should be a dynamic and smart configuration, capable to adapt and update itself automatically and continuously by following the interaction with the user.</p><blockquote>The more the user interacts with the smart system, the more the system learns and adapts its configuration accordingly.</blockquote><p>A good example is ‘The Grid’, an AI-driven system to generate websites, where the layout and the look and feel of the page adapts automatically to the content inserted by the user. In general, a smart system should be able to make the most user data and contextual information to make the first use of the service even more simple, personalised and relevant for the user.</p><h4>Keep the user always in control.</h4><p><strong>Data, settings and preferences management is a central part of a smart experience. </strong>How user data and preferences are collected and used, and how the user is able to control and manage these data and processes? This is one of the most important and controversial point of a smart user experience. One implication of a smart services is the lack of direct control: outcomes and interactions are managed automatically by a system, the user can change the inputs, but the relationship between inputs and outputs is non-linear, complex to learn and the user may not be able to repeat the same conditions twice. This goes completely against any traditional HCI approach that requires a simple mental model that enables a user to predict the reaction of a system to his actions and to be able to repeat it consistently.</p><p>If data and preferences are crucial for smart service outcomes, the user should have the control on the inputs and the outputs of the system, and be able to manage data, settings, preferences and rules that feed the system. Today most of these features are managed with settings, control panels, on-boarding processes, preferences, questionnaires and many other similar features that usually are pretty hidden, boring and secondary for the user experience of many applications and digital services. For example, we all know very well how is complicated to manage the privacy settings in Facebook or the notification system of an internet banking.</p><p>These features are crucial not only for transparency and for the sake of personal data and privacy, but also to give the user a real control on the service and its outcomes and to make it more personalised, useful and relevant. The user should able to fine tune and update the system configuration continuously, making key settings and preferences easily accessible almost at every step of the experience. The user should be able to see what is behind every automatic decision of the system (“<em>you see this because…</em>” - “<em>this happens because…</em>”), in order to learn it and manage it and change the service result at will. Finally, the user should be able to reset completely the settings and program again the system. Configuration features and control panels should be designed properly in order to be more accessible, comprehensible and usable. To provide the user with a good overview of the potential consequences of different settings and choices, the effect of a configuration change on the service outcome should be visible and understandable, also in advance, possibily with graphic simulations and visualisations.</p><h3>Conclusions</h3><p>Today many smart systems and experiments are still driven mainly by technology. This is a natural step for any emerging technology before getting mainstream. But the “smartification” of objects is not enough to engage the user and make successful commercial applications. Beyond single connected gadgets and the technical framework, the user value lies in the consistency of the smart platform and in its capability to deliver useful services grounded on actual needs and goals.</p><p>A well designed user experience is the key for the success of any hyperconnected smart technology. The more connected devices and the technology blend in around us the more a good smart user experience is crucial for service adoption and efficacy. As Mike Kuniavsky writes, we need a new integrated design approach for platforms and services that combines interaction design, industrial design and service design to create smart experiences centred on the users and their real needs and desires.</p><p>A special thanks to <a href="https://medium.com/@thomas_thinks">Thomas Sutton</a>.</p><p>References. <em>This post is the result of many readings, conversations and discussions (and experiments with smart gadgets) occured over the last two years. The following references are just the most relevant and the most recent, but surely I am missing many others.</em></p><p>(1) K. Worden, W.A. Bullough, J. Haywood, <em>Smart Technologies</em>, 2003, World Scientific Publishing Co.</p><p>(2) Michael E. Porter, James E. Heppelmann, <em>How Smart, Connected Products Are Transforming Competition</em>, HBR, 2014 <a href="https://hbr.org/2014/11/how-smart-connected-products-are-transforming-competition">https://hbr.org/2014/11/how-smart-connected-products-are-transforming-competition</a></p><p>(3) Tim Worstall, Eric Schmidt’s Quite Right The Internet Will Disappear; All Technologies Do As They Mature, Forbes, 2015 http://www.forbes.com/sites/timworstall/2015/01/24/eric-schmidts-quite-right-the-internet-will-disappear-all-technologies-do-as-they-mature/</p><p>(4) Mark Weiser, The Computer for the 21st Century, 1991</p><p>(5) Justin Zalewski, Interaction Design within the Internet of Things, 2014 <a href="http://www.kaplusa.com/interaction-design-within-the-internet-of-things/">http://www.kaplusa.com/interaction-design-within-the-internet-of-things/</a></p><p>(6) Mary Treseler, How is UX for IoT different?, O’Reilly Radar, 2014 <a href="http://radar.oreilly.com/2014/11/how-is-ux-for-iot-different.html">http://radar.oreilly.com/2014/11/how-is-ux-for-iot-different.html</a></p><p>(7) + Rehabstudio <a href="http://www.internetofuselessthings.io/">http://www.internetofuselessthings.io</a></p><p>(8) Azure Yang, Simone Rebaudengo, <em>When Objects Talk Back, </em>Design Mind, 2014 <a href="http://designmind.frogdesign.com/2014/10/when-objects-talk-back/">http://designmind.frogdesign.com/2014/10/when-objects-talk-back/</a></p><p>Other references:</p><p>Steven Hoober and Mudassir Azeemi, <em>People, Ubiquity, and the Internet of Things: The New Mobile Context</em>, UX Matters 2014 <a href="http://www.uxmatters.com/mt/archives/2014/01/people-ubiquity-and-the-internet-of-things-the-new-mobile-context.php">http://www.uxmatters.com/mt/archives/2014/01/people-ubiquity-and-the-internet-of-things-the-new-mobile-context.php</a></p><p>Claro Partners, <em>The Internet of Things calls for a new approach to design</em>, 2014 <a href="http://www.claropartners.com/the-internet-of-things-calls-for-a-new-approach-to-design-and-why-sticking-a-sensor-on-it-isnt-enough/">http://www.claropartners.com/the-internet-of-things-calls-for-a-new-approach-to-design-and-why-sticking-a-sensor-on-it-isnt-enough/</a></p><p>Antonio Regalado, <em>Business Adapts to a New Style of Computer</em>, Technology Review, 2014 <a href="http://www.technologyreview.com/news/527356/business-adapts-to-a-new-style-of-computer/">http://www.technologyreview.com/news/527356/business-adapts-to-a-new-style-of-computer/</a></p><p>Dave Grey, Everything is a service, 2011 <a href="https://www.sprinklr.com/social-scale-blog/everything-is-a-service/">https://www.sprinklr.com/social-scale-blog/everything-is-a-service/</a></p><p>Michael Kuniavsky, Smart Things: Ubiquitous Computing User Experience Design, 2010.</p><p>Michael Kuniavsky, A pattern for user experience design in predictive analytics <a href="http://www.wired.com/campaigns/theconnectivehub/?p=390">http://www.wired.com/campaigns/theconnectivehub/?p=390</a></p><p>Steve Denning, CES: A User’s Guide To The New Economy, Forbes, 2015 <a href="http://www.forbes.com/sites/stevedenning/2015/01/14/why-the-consumer-electronics-show-is-dying/">http://www.forbes.com/sites/stevedenning/2015/01/14/why-the-consumer-electronics-show-is-dying/</a></p><p>Tom Simonite, 2014 <em>in Computing: Breakthroughs in Artificial Intelligence</em>, Technology Review, 2014 <a href="http://www.technologyreview.com/news/533686/2014-in-computing-breakthroughs-in-artificial-intelligence/">http://www.technologyreview.com/news/533686/2014-in-computing-breakthroughs-in-artificial-intelligence/</a></p><p>Lelylan Blog, We need to improve the Internet of Things User Experience, 2014 <a href="https://medium.com/@lelylan/web-apps-for-the-internet-of-things-329fa683303">https://medium.com/@lelylan/web-apps-for-the-internet-of-things-329fa683303</a></p><p>Thomas Sutton, Invisible Applications, 2015 <a href="https://medium.com/@thomas_thinks/invisible-applications-10501f6cfa5">https://medium.com/@thomas_thinks/invisible-applications-10501f6cfa5</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a6e675b414ec" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[UX Design: the user is not enough]]></title>
            <link>https://medium.com/@lowresolution/ux-design-the-user-is-not-enough-c90d595eb237?source=rss-693ebd7d07c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/c90d595eb237</guid>
            <dc:creator><![CDATA[Gianluca Brugnoli]]></dc:creator>
            <pubDate>Sat, 02 Nov 2013 15:00:14 GMT</pubDate>
            <atom:updated>2016-02-14T21:17:43.315Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*tY_86-buvMDR8r79.jpeg" /><figcaption>Vespa Piaggio 1946 BLUEPRINT</figcaption></figure><h4>For a successful project the focus on the user is not enough: Design leaders are also product experts who deeply understand the business challenges of the project.</h4><p>Historically the practice and culture of Design have been defined by what has to be designed in a project. A car, a dress, a furniture, a domestic appliance, a computer and so on. In the era of product design, designers were mainly segmented on vertical industries and focused on the business and the technical features of each specific product, and the end user was seen as an inconvenience or a dumb consumer to be educated.</p><p>The rise of user-centered methodologies, practices and tools changed everything: the center of the project shifted from the product and the technical features to the users and their experience, breaking the vertical industry silos and bringing many fundamental and positive innovations in the way products and systems today are conceived and designed.</p><p>The user-centered revolution has been possible mainly rebuilding the creative and design process and developing a new generation of design tools. The <a href="http://www.dubberly.com/concept-maps/creative-process.html">design process</a> has been mapped, structured, re-organized and reordered in stages and modules, from research to implementation, each one with its own set of specific goals, activities and deliverables. Once the design process was unstructured and unpredictable, now it can be described, organized, planned and even replicated in detail, step by step, deliverable by deliverable. Alongside lots of new <a href="http://www.servicedesigntools.org/">design tools</a> and techniques have been developed and combined into the design process to analyze, map and unravel the complexity of the user, of the systems and of the product mix.</p><p>Unquestionably these innovations have extended the design culture and practice way beyond its traditional range to new domains and challenges, making Design even more cross-disciplinary and complex.</p><p>Today designers and design thinkers are taking on the challenges of designing experiences, services and strategies, mostly following the same multidisciplinary and holistic approach to the creative process developed for user-centered design. These tools and techniques are applied to broader and more complicated business problems. Apparently the loop is closed: complex product ecosystems and service platforms are now designed putting the user in the center and connecting the dots of the multiple touch-points of the user experience. Nonetheless one of the key problem of UX Design and Service Design is that many designers often seem to be engrossed with the design process and its tools, without paying sufficient attention to the specific business and product challenges that could make the project successful.</p><p>Looking over the most recurring points of the current User Experience and Service Design debate, mostly it is about defining and structuring the design process and its deliverables, the workshop formats, refining the design tools, the design phases and so on. This is fine, but the attention paid to understand an define the real challenges of the product and service that has to be designed, and what makes it really different and innovative for an organization, is often little.</p><p>There is no question that user experience should be the center of everything, and that understanding and empathizing with the users is a crucial part of the success of an innovation project. Nonetheless designing a service or an experience for banking, for retail or for the healthcare sector is a completely different problem, although the design process, the activities and the design tools can be the same.</p><p>When it comes to prove the business value of User Experience Design and thinking, the concern of the UX community is about how to communicate and explain the benefits of the various parts of the UX Design approach to clients and business stakeholders. The confidence in the appropriateness of the approach and of the design process is so big, that for many designers the best way to show the value of their work is to persuade business stakeholders of the benefits of investing in it and making them realize the negative impact of not including a strategic design thinking, beyond the tactical design execution.</p><p>This is not wrong, of course, but the quality of communication and packaging are just a part of the problem. The problem of the business value of UX Design is not only communicating more efficiently the value of the design process and of its structure and outcomes, but to make these parts really relevant for the specific business goals of the project.</p><p>Everybody realizes that there is gap between design and business, but the question is not that business stakeholders don’t get UX Design: sometimes UX professionals fail to master the specific industry challenge behind the project, and sometimes they seem not even interested in solving it. As a consequence, many clients and corporate stakeholders see UX and Service designers as generalists who want to deal with complicated challenges without the appropriate vertical knowledge and competences. Moreover many stakeholders struggle to understand the value delivered by a workshop or a design phase, even when a great value is actually delivered.</p><p>For a successful project, user-centered design processes and tools are definitely necessary, but not sufficient. If User Experience Design wants to get to the strategic level, designers should be also product and industry experts, able to understand and master the strategic business background and objectives and the real challenges of the specific product and service they are designing.</p><p>Great designers deliver design solutions to business problems. It’s not only a matter of understanding the users and the context of the experience, and it’s not only about problem solving. Being experts of the design process is a crucial skill for a designer, but it’s not enough. The process is only a tool, a means to an end. Design leaders are able to get into the core of the specific business problem and goals of the project, and then combine and coordinate the various design activities and tools structuring the design process consequently.</p><p>The true business value of UX strategy does not lie only in a well structured design process, but in a design process properly structured on the actual business challenges to solve.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c90d595eb237" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>