What does it mean to weave intelligence into an interface? How do we add AI to digital products in ways that are at once meaningful and responsible? What is the grain of this new design material, and what are its strengths and weaknesses? What are the opportunities and the risks, and how do we balance them?
Hey, those are hard questions! We’re finding ways to answer them with a practice we call Sentient Design.
Sentient Design is the already-here future of intelligent interfaces: experiences that feel almost self-aware in their response to user needs. Sentient Design moves past static info and presentation to embrace UX as a radically adaptive story. These experiences are conceived and compiled in real time based on your intent in the moment—AI-mediated experiences that adapt to people, instead of forcing the reverse.
Sentient Design describes not only the form of this new user experience but also a framework and a philosophy for working with machine intelligence as design material. As our interfaces become more mindful, so must designers.
And! Sentient Design is also a book —or will be soon. Rosenfeld Media will publish Sentient Design: AI and the New Digital Experience by Josh Clark with Veronika Kindred. Look out for it late this year or early 2025. We’ll get it into your hands as soon as we can!
This is so much more (and in exciting ways, so much less) than chatbots, text prompts, and systems that can write, speak, or draw. In a nutshell, Sentient Design is all about the new opportunities to add intelligence to our interfaces (sometimes just a dash, other times a heaping spoonful). Some of this is so familiar that you might consider it nothing special (content recommendation or prediction), but there are lots of surprising new possibilities here, too. It’s all part of our kit as designers now—and especially powerful in combination. Sentient Design explores all of it.
This essay is a high-level introduction to Sentient Design’s core concepts. To dive straight into the details, check out Josh’s lively Sentient Design talk, which introduces lots of examples and design principles. The talk is available in video, audio, slides, and prose.
You can also keep an eye on our series of Sentient Design articles and talks for detailed examples, design patterns, and strategies. Sign up for our newsletter A Little Big Medium to keep up with all of it, including news about the book.
Why Sentient Design Matters
All around us, there’s something that you feel more than see, an undercurrent that swirls through your everyday. It’s the influence of a million different algorithms supporting, nudging, cajoling… quietly working beneath the surface of nearly every digital experience—and many physical ones, too.
This eddy and flow of machine intelligence infuses everything, serving us in countless ways, often invisibly. More recently, these systems have become more visible partners; everyday folks now interact directly with AI through chat or other experiences. Much of the time, all of this is for the good. Steady leaps in machine learning and AI have made possible a slew of near-magical products and services that ease and improve our lives in ways large and small. The opportunities are everywhere and growing.
But so are the risks and failures. Flaws and mistakes range from the comically trivial (looking at you, autocorrect) to the deadly serious. AI systems have wrecked lives with biased prison sentencing, bungled medical diagnoses, and plane-crashing autopilots. At an even larger scale, we’ve seen AI undermine faith in entire political systems by automating the broadcast of propaganda and hate speech with both a scale and targeted focus that are unprecedented. Others use machine learning to manipulate or monitor us in less profound but equally cynical ways for advertising, surge pricing, surveillance, and more.
These are not entirely or even mostly technology problems. Instead, they reflect a still-developing understanding of how to put AI to work in ways that are not only useful and valuable, but also respectful and responsible.
Data scientists and algorithm engineers have revealed the possible; now design and other fields can join to shape that potential—and reveal the meaningful.
For all of its pervasiveness, our use of AI and machine learning is still in its awkward adolescence. Flashes of maturity mingle with stunning lapses of judgment, and boastful confidence gives way to naive misunderstanding—and sometimes dangerous behavior. These adolescent systems are still gawky and undisciplined, but the same could be said about our approach to building them.
Sentient Design aims to help all of us move through our growing pains and into this next chapter of experience design. That means helping product organizations grow, too. How should makers and managers craft products and services in the era of the algorithm? What shape should those experiences take? How can you take advantage of the powerful opportunities of AI while guarding against its risks and failures? What are the meaningful problems you can solve and new questions you can ask with these tools in your kit?
These are design challenges, not technical. Data scientists and algorithm engineers have revealed the possible; now design and other fields can join to shape that potential—and reveal the meaningful.
What are Sentient Design experiences made of?
Sentient Design refers to intelligent interfaces that are aware of context and intent so that they can be radically adaptive to user needs in the moment. Those are the fundamental attributes of Sentient Design experiences: aware and radically adaptive.
Those core attributes are supported by a handful of common characteristics that inform the manner and interaction of most Sentient Design experiences:
- Collaborative. The system is an active (often proactive) partner throughout the user journey, often with independent ability to perform tasks on your behalf.
- Multimodal. The system works across channels and media, going beyond traditional interfaces to speech, text, touch, physical gesture, etc.
- Continuous and ambient. The interface is available when it can be helpful and quiet or invisible when it can’t.
- Deferential. The system suggests instead of imposing ; it defers to user goals and preferences. It offer signals, not answers.
Note that “making stuff” is not included in this list—not explicitly, at least. Writing text, making images, or generating code might all be the means or even the desired outcomes of Sentient Design experiences—but they’re not defining characteristics.
The fundamentals of Sentient Design aren’t new. Some aspects of Sentient Design have been around for years or decades—long enough that they’ve become familiar design patterns. We encounter automated services hundreds of times per day that deliver bespoke experiences based on automated recommendation or prediction.
The commonality is that these services are aware and adaptive in presentation. Streaming music services are aware of our preferences and history, as well as what’s new and popular in the mainstream, and adapt content and presentation; the result is a playlist (or several!) that suits our tastes, or maybe just the moment. Similarly, predictive keyboards are aware of what you’re typing, along with the historical probability of the most likely next word or phrase; they adapt the keyboard in anticipation of what we’ll type next.
These kinds of just-in-time content presentation are familiar enough to seem commonplace, but they remain part of the Sentient Design toolkit alongside the fancy new stuff like generative AI and large language models (LLMs). As the underlying tools become more capable, awareness and adaptation grow more capable too. For example, AI-powered assistants can understand any question and come back with an appropriate (and sometimes even correct!) response. Talk about radically adaptive: you can craft a whole meandering conversation that is absolutely one of a kind. It’s an experience that totally bends to your want and need in the moment—indeed, it’s invented for that moment.
What’s also new is that Sentient Design experiences anticipate our needs based on models of what we have done in the past, individually or collectively. They propose paths forward, show us new possibilities, caution us about blind spots, or maybe just take care of simple tasks. These intelligent interfaces can also come to us instead of the reverse, shape-shifting into the form and interaction that is most convenient. In some cases, that may mean reformatting content or structure on a screen, or changing the UI metaphor entirely. More ambitiously, the AI-powered ghost in the machine may flit into a different machine entirely—shifting the point of interaction on the fly from screen to earbud to hand tool to appliance to car cockpit and back again.
New experiences, new perspective
This isn’t exactly ye old website anymore! While Sentient Design has an important place in traditional interface design, too, it introduces a very different experience to conceive and manage than the relatively static experiences of forms and dialog boxes that many digital designers have known so far.
In this new kind of experience, everything can potentially adapt on the fly. The content, the format, the mode of interaction, the point of contact, the tone and aesthetic, even the very goals of the system… All of these things can shift and adapt in a new kind of nimble collaboration with the people it serves. The tools and design patterns are here for us to use now.
It’s exciting and delicate and maybe even a little weird.
It is not as weird as full-blown consciousness. We’re not talking about machines with emotional feeling, moral consideration, or any of the hallmarks we might ascribe to a fully sentient being. The “sentient” in Sentient Design describes a combination that is far more modest but still powerful: awareness, interpretation, agency, and adaptation. Machine learning and AI can already enable all of these attributes in forms that range from simple to sophisticated.
This is a continuum. Think of it as a dial that you can turn: from simple utilities that add sparks of helpful intelligence to humble web forms to more capable companions like agents or assistants. Those experiences are comfortably attainable with Sentient Design and today’s technologies.
The opportunity is to begin thinking of machine-learning applications in the same way that you think of using JavaScript to layer interactivity into a webpage. Or how you use responsive design techniques to add flexibility and accessibility to an experience. Just like we’ve absorbed those techniques into our daily practice, we can do the same with Sentient Design. Machine learning features are just another new technique and applied technology, this time to add intelligence and awareness to our interfaces.
We can add that spark of intelligence anywhere it can be helpful. Apply Sentient Design’s awareness of context and intent, using things like recommendation, prediction, classification, clustering, and generation to elevate humble web forms or create entirely new kinds of experiences. What data do you have that can anticipate next steps and help people move more quickly through challenging or mundane tasks?
Grounded in solid UX principles
This is a lot—and it comes with new challenges and dangers. We invite trouble when we abdicate too much responsibility to intelligent systems, or when we misrepresent their statistical signals as facts or absolutes. It takes new perspective and new technique to realize the opportunities of AI while contending with its risks.
Instead of designing for success—for the happy path—we have to do more to anticipate failure and uncertainty. Sentient Design experiences allow AI and machine learning to mediate experiences in ways that are new and that sometimes take the designer out of the loop. We have to anticipate how and where the system is unreliable—where the system will be weird and where the human will be weird.
We invite trouble when we abdicate too much responsibility to intelligent systems.
Sentient Design introduces several principles for defensive design. How do we anticipate and manage the imperfect or frankly bizarre results that AI often delivers? How do we design the experience to set expectations and channel behavior in a way that best matches the system’s ability? And perhaps most important of all: how can Sentient Design experiences engage critical thought, or amplify human judgment and agency, instead of replacing them?
Sentient Design is about AI… but also not. While AI and machine learning are the enabling technologies, the goal is not to “make AI products.” Sentient Design lets us deliver meaningful human outcomes in ways that haven’t been possible until now. The intended outcomes themselves may not change—our creations should remain focused on human goals—but Sentient Design can collapse the effort it takes for people to realize those outcomes, and may even make some outcomes plausible, affordable, or attractive for the first time. Along the way, Sentient Design techniques can make experiences more valuable, more intuitive, and often more fun. That’s the real opportunity; the technology is only the means to the end.
How to learn more
We’re working hard to share the details of how all of this works in practice. Here are some resources to help you stay current.
Watch or listen. Check out Josh’s Sentient Design talk, which introduces lots of Sentient Design examples and design principles. The talk is available in video, audio, slides, and prose.
Read up. Our ever-growing series of Sentient Design articles and talks offer perspectives, examples, and design patterns. Sign up for our newsletter A Little Big Medium to keep up with all of it, including news about the book.
Workshops and Sentient Design sprints. Big Medium offers educational workshops about Sentient Design and the UX of AI. We also work with clients to conduct Sentient Design sprints—sessions for product designers to identify, frame, and validate applications of machine intelligence. Get in touch for details.
Product design engagements. Big Medium does lots of product design, and Sentient Design is a big part of our toolkit. If you’re considering AI-powered products or features, we can help. Get in touch.
