Big Medium is starting the year strong with a slew of “design for what’s next” engagements—and together they send interesting signals for the year ahead. Our clients are pursuing ambitious, forward-looking efforts, and we’re excited to help realize them. Just a few of our current projects:
- We’re guiding product and UX strategy for a medical software company to create radically adaptive, AI-mediated experiences that will ease and transform doctor workflows.
- We’re creating new tools and processes for one of the world’s largest companies to deliver meaningful AI-assisted design solutions for improved quality and speed.
- We’re reimagining brand and messaging architecture for one of America’s leading media companies to meet a transforming media landscape.
- We’re elevating an already mature, successful design system to support cross-platform AI-powered applications for an enterprise category leader.
- We’re writing the Sentient Design book for Rosenfeld Media to describe how to build a new generation of intelligent interfaces.
- We’re teaching Sentient Design workshops to internal teams across a range of companies as designers adopt AI as a new material for product design.
These projects are exciting on their own, but what’s got me tingling is how the highlighted tidbits of those forward-looking efforts signal broader industry trends I’m seeing for the new year.
I’m in constant conversation with digital leaders about their strategic and operational hopes and struggles. Last year, AI generated lots of uncertainty and a ton of tentative pilot-project explorations. Perspectives and projects are changing now, and as we anticipate and move with these changes, that means a shift in emphasis and approach for Big Medium, too. Here’s what it adds up to for the new year:
- AI shifts from tool to design material
- Design shifts from consolidation to innovation
- AI learns design systems
- Generative AI becomes boring
- AI hype continues to distract
- New challenges need new perspective
AI shifts from tool to design material
The AI focus for most companies has so far targeted process tooling or product features that focus on productivity and function: How does AI deliver efficiencies? What can generative AI make or do? This mindset has privileged feats of engineering over feats of experience. Product designers have focused more on how AI might change their process than how it could elevate the products they make.
That will change this year. Instead of treating AI as a tool, a function, or even as an enabling technology, we’re seeing design leaders start to explore it as a material for new product experiences. What becomes possible when you weave intelligence into the fabric of a digital interface instead of bolting on some sparkles and a chatbot? You get AI-mediated experiences conceived and compiled in real time based on your intent in the moment. These are radically adaptive interfaces that adapt to people instead of forcing the reverse.
Instead of cost-cutting efficiencies, this approach emphasizes how AI can increase product value by transforming experience. In our Sentient Design practice, we’ve identified 14 new experience patterns that intelligent interfaces enable, and we’re helping our clients deliver them now.
This goes way beyond chatbots, and it’s exciting work, exploring and establishing new kinds of intelligent interfaces. Sometimes, this makes for dramatically new interaction models, but other times, it can be way more subtle—casual intelligence sprinkled into traditional interfaces like web forms.
Big Medium is making this next chapter of interaction and product design a core focus in 2025, and we’re excited to help the industry turn this corner, too. We’re doing that in a few different ways: intelligent interfaces are central to a growing portion of our product design work; Veronika Kindred and I are writing the Sentient Design book for Rosenfeld Media; we’re giving talks at SXSW and conferences around the world; and we’re teaching Sentient Design workshops to product teams who want to learn how to make what’s next.
Speaking of that appetite for “what’s next,” that’s another strong industry shift we’re seeing for product design and strategy…
Design shifts from consolidation to innovation
We’re emerging from a long period of design retrenchment. The last 15 years has been marked by companies bringing design in-house—and learning how to metabolize that function to make creative process work inside company structures. That real and important need gave rise to the design system discipline over the past decade. Design systems represent the systemization and adoption of best practices. They’re the best way to get everyone to work from the same playbook.
But design systems are about curation, not invention. Their whole job is to consolidate, not innovate. The real innovation was in creating the design system discipline in the first place—the assets, technique, and process to make them go. That’s been an area where Big Medium has contributed a ton for many years. My longtime partner Brad Frost developed the Atomic Design methodology across our many projects together. We contributed directly to the success of hundreds of design systems—and indirectly to hundreds of thousands.
But now… mission accomplished. Design system best practices have settled; the discipline has become established; and having a design system is now considered the baseline. (That doesn’t mean it’s easy. Brad continues to do the important work of helping teams put these established practices to effective work. Brad has moved on to focus exclusively on delivering this content and education at scale. Check out Subatomic: The Complete Guide to Design Tokens, Brad’s new course with Ian Frost—so great!)
Meanwhile, Big Medium is focused on the next wave. I’m seeing a fresh appetite—and a real need—for product design innovation. Creative momentum is shifting from the consolidation of best practices to the invention of new ones. I haven’t felt this level of curiosity, urgency, and enthusiasm to explore new digital experiences since mobile changed everything. AI-mediated experiences are that new thing, and nurturing its innovation and mastery is the big focus of Big Medium’s product design and strategy work right now. We’re excited to help lead that exploration at both the client and industry levels.
Creative momentum is shifting from the consolidation of best practices to the invention of new ones.
But this is cool: the last chapter informs the next. All that design-system work creates the foundation to enable AI-mediated interfaces.
AI learns design systems
Big Medium’s design system practice is shifting to focus on helping mature teams to capitalize on the interplay of design system and machine intelligence. Part of that is the productivity angle of how AI can better enable the maintenance and adoption of design systems—we’ve been doing some impactful work on that with our clients. But the bit that’s even more exciting and novel is the role that design systems play in creating radically adaptive, intelligent interfaces.
The essential job of a design system is to create a library of solved UI problems organized around common language and purpose. Everyone is on the same page: when you see this UI problem, you know to reach for that UI solution. When AI-powered systems become part of the team, they can make use of that design system know-how, too.
One aspect of Sentient Design is interfaces that morph according to immediate context. Modern chat experiences can field any question and then take the conversation in any direction the user cares to take. What happens when other interaction models take on the same open-ended flexibility? You get things like Google’s bespoke UI or Salesforce’s generative canvas, where content and interface elements are chosen on the fly—but always from the design system.
More than just a new experience, it suggests a new role for designers and design systems, too. When we allow machine intelligence to share real-time UI decisions, the designer’s job shifts to designing the rules and materials that govern those decisions. That means building robust design systems that can be flexibly composed, developing clear principles for when and how experiences should adapt, and establishing guardrails to ensure coherence even as presentations shift. The experiences themselves may be ephemeral—built for the moment—but the foundations or generating them must be even more solid than traditional interfaces.
When AI-powered systems become part of the team, they can make use of the design system, too.
We’re already helping product design teams build these forward-looking experiences—in reliable ways that don’t turn the interface into a robot fever dream. And we’re helping mature design system teams explore how to support radically adaptive interfaces with their UI patterns. I expect much more of this in 2025.
Generative AI becomes boring
This is already happening as last year’s magic turns into this year’s everyday software. Systems that can hold conversations on any imaginable topic (!!) are just… taken for granted. Image generation, writing, summarization, natural language understanding—all of this stuff is already getting baked into everyday operating systems, fast becoming as expected and commonplace as spam filters, predictive keyboards, or recommendation systems before them.
This has been the path of AI for the last 75 years—every breakthrough quickly becomes “just software,” and AI gets redefined as something better/smarter. AI has always been a moving target hovering just out of reach, but the broader category of machine intelligence has kept chugging away, putting AI’s castoffs to work on sober, workaday tasks. It’s the algorithms and machine learning that power the software we use every day. While machine intelligence enables AI’s sensational new experiences, it also enables a slew of boring but important features and business processes. “Somehow machine learning is both rocket science and tractors,” says analyst Benedict Evans.
The new banality of generative AI is a good thing, because it means that we can treat this design material with calmer rationality. We can understand what it’s good at and bad at, find its proper place in our products and workflows—and leave it aside when it’s not fit for purpose.
At Big Medium, we’ve always focused on “teaching teams to fish” in our client engagements. As we work alongside client teams, our goal is to make our knowledge their knowledge. In 2025, I expect a big part of that will be helping our partners adopt AI as part of their everyday toolkit without fuss or fanfare. It’s just software.
AI hype continues to distract
If AI developments stopped right now—if frontier models got no better and advancements froze—we would still have plenty of material to develop years and years of new product experiences. The pace of technology always outpaces the ability of organizations to adopt it, and the astonishing speed of AI developments makes it especially true in this area. There’s a lot to process here and so much yet to create with the actual concrete tools of machine intelligence available today—with all their capabilities and shortcomings.
And yet the constant drumbeat from big tech and the companies behind the big foundation models is: wait til you see what’s behind the curtain. It’s always the thing that’s about to arrive. There’s a ton of money at stake in keeping anxiety aroused and anticipation heightened—billions and billions being invested in promises to make the next generation of machine intelligence even better. The dangled promise is AGI—artificial general intelligence—the threshold when machine intelligence can do any general task better than human intelligence.
I have no idea whether and when AGI will happen. People way more informed than I about the science of this stuff disagree strongly about the timing or even the plausibility. We’ll see.
But whatever happens next, the hype and distraction about what might be tends to distract from what’s possible now. Language and expectations race far ahead of what’s delivered. At Big Medium, we’re very pragmatic in our approach to machine intelligence. We keep an eye on what’s emerging, but we’re very focused on what’s practicable, affordable, and ethical within the current set of tools.
I share the industry’s current fascination with the promise of AI agents—self-driving software that can set goals, plan and execute the steps to achieve them, and then recognize when the goal is complete. But I also don’t think 2025 will meet the breathless expectations that agents will handle all manner of open-ended tasks. 2025 will not be the year of autonomous, general-purpose agents. The oft-discussed software agent that can plan and book your vacation will remain elusive.
But! I do expect that agents will begin to handle more and more tasks in narrow domains within carefully scoped systems. That’s all within the capabilities of current tools—effective agents are already available for research and for software development. A big part of our work in both process and product is moving toward agentic/agentive features and applications in similarly scoped domains. We’re helping our clients transition from traditional automated tools to the first stage of AI-assisted agents that bring context and awareness to narrow tasks.
The hype and distraction about what might be tends to distract from what’s possible now.
We need to proceed with care and pragmatism. The work and the challenge for practitioners in 2025 is to develop the meaningful stuff that’s possible right now without overreaching what current tools can do—or falling into a hype-fueled hypnosis about what might be possible tomorrow.
New challenges need new perspective
The media and digital landscapes are changing. There are new ways to tell stories and share information through interactions that weren’t possible just a few years ago. We can convert content between style, tone, and medium—PDFs to podcasts—in minutes or seconds. We can generate both meaningful content and useless slop faster than ever. We can surface insights and hallucinations with astonishing ease, but often find it hard to tell the difference. What do we do with this?
AI has put companies, teams, and individuals on an existential roller coaster ride for the last couple of years—what does it mean to who we are and what we do? In what areas will our skills as individuals (and capabilities as companies) become even more important this year? Where should we acknowledge that AI is already better at some of our traditional tasks? What are the meaningful new offerings that emerge when we combine forces with machine intelligence? How does that change the way we are perceived and what we’re capable of?
For leaders, the work in 2025 is to develop clear-eyed perspective about this—and to craft crisp internal and external messaging to communicate it. We’re already helping our clients do this at the project, team, and enterprise levels, but I expect that work to broaden and deepen in 2025. That means developing new design principles around next-generation product work. It means developing new professional development programs to help teams adopt the skills and perspectives they need for their work. And it means developing new positioning and brand strategies at the organization level to help the world understand how you’re evolving. We’re doing all of this in our client work now.
2025 will be a year of transition. This won’t all be easy or comfortable; it means changing old habits and sometimes sense of self. But I believe this transition will also unlock a year of energetic invention—one that will be as fun and rewarding as it is challenging. Here at Big Medium, I’m excited to see that our current slate of client projects is already starting us quickly on that course. I’m even more excited about the story that those projects tell me about the year ahead for all of us.
Happy New Year, friends. There’s lots to do. Let’s get after it.
Is your team wrestling with what’s next for product strategy, design, or process? We can help! Get in touch to learn more about our engagements for product strategy, design/development, and Sentient Design workshops.
