What’s next for design teams?

AI tools have changed so much for individual designers. The second-order effect is on teams. What is the role of a design team in a company, now that individuals from all backgrounds are able to bring their own designs to life?

I’ve been talking with folks on many types of teams, and have started to see some interesting patterns and opportunities for design teams more broadly.

At a high level, the biggest opportunities are around not just producing design, but multiplying it. Teams need to understand that everyone in an organization can ship design work now, and focus their efforts on helping that grow while simultaneously raising the design standard to a higher level.

1. Pay attention, as an organization

I wrote recently about how people dedicated to paying attention to design will remain an important part of successful product teams. The same thing is true at the organization level. Who is paying attention to company-level design issues?

• Is the brand evolving as desired, and are products evolving to meet it?
• How are leaders talking about and considering design in decisions?
• Are products connected to each other in thoughtful, elegant ways?
• Is anyone making sure the company has a vision for the future?

Promoting and maintaining attention to design at the company level takes education, consistent pressure, and deep knowledge of great design. Design teams should be the source of these efforts, and measured by their success.

2. Keep the bar high, systematically

It’s wonderful that people from every role and background can now visualize and contribute their ideas. But not all ideas, and not all designs, should be immediately merged and pushed live. Endless uncoordinated changes lead to disjointed, clumsy experiences that frustrate customers.

And yet no one wants to be the bad cop blocking their colleagues all day. Many designers I spoke to referenced how their role in maintaining design quality was exhausting as they were asked to review dozens of UI changes each day from all corners of the company.

One solution is redefining the “design system” one more time. An AI design system lives in code and is readable by agents. When that’s in place, designers can build agents, skills, and evals on top of it to be the first reviewer on proposed UI changes. Those agents can look at a well-documented design system and give rational, emotionally neutral feedback on known design system conflicts, reserving human review for higher-level discussions.

3. Be the inspiration

One of the most fun and effective practices my design teams have done at multiple companies is to bring regular, diverse, and thought-provoking inspiration to product and engineering teams. Often this takes the form of a weekly newsletter with snippets of work, things we’ve learned, and external products, ideas, and resources we’ve discovered.

Being the source of this inspiration is a powerful way to spark new ideas, conversations, and influence among the team. It also takes a lot of time to gather material from your team and everything they’re drawing on.

With AI doing the synthesis, this sort of practice is easier than ever. You can set up an agent to gather these resources for you, and simply choose which items to share out. Agents can look further and broader than you ever could, and bring you things that are truly inspirational.

For a design team, this might be an evolution of a content design practice, or a design ops one. In both cases the “customer” will often be an internal one.

4. Create tools and templates

In all of these cases, while doing the work yourself is good, it’s even better to make that work possible for others. You can sometimes just as easily build a tool for others to gather inspiration, prompt an exploration, or map a user journey. Stop thinking of your team as just designers of objects; think of them as designers of processes.

Design teams should have people dedicated to building design tools and templates for others, not just designing products directly.

5. Actually design “the experience”

The title “UX designer” has always been a bit of a reach. Sure, we talk about “designing the experience”, but most of our time has been spent moving pixels between rectangles.

With AI doing the pixel pushing, we’re free to consider the broader experience people have with our products, working at the level of customer journeys, emotional state, and jobs to be done. Blending those solutions with the detail and care for the UI is a powerful combination.

It also keeps us focused on the human with a need, not the display on their screen. Human designers will maintain better intuition about human emotions and desires than an LLM for the foreseeable future, and be a better judge of potential solutions.

Design teams can organize around and focus on customer problems rather than product features. Their impact should be seen in cross-team collaboration and elegant cross-product experiences.

6. Frame the problems

The old UX joke: “How many designers does it take to change a light bulb? Well… does it have to be a light bulb?”

Designers are infamous for reinterpreting and reframing product requests. It’s a habit we just can’t help, and it’s equally valuable and frustrating to our colleagues. In fact, when I trained my design agent on my own behaviors, I started getting frustrated with it too! It’s never easy or comfortable to be challenged on your own ideas.

Yet this is something unique to our role and perspective, and powerful. Designers are skilled at seeing problems through different lenses, comparing and contrasting them, and finding a framing that is elegant and clear.

Design teams should look across the problems they’re being asked to solve, and build a better way of thinking about them. That might mean reorganizing the team around new experience pillars, or proposing a design evolution that guides multiple efforts into one unified direction. A product strategy is also a design, of course.

Now that design and prototyping is more accessible than ever before, the role of a design team is shifting from producing design to multiplying it: focusing attention, setting the bar, building the tools, and framing the problems everyone else is now empowered to solve.

The language of design

One of the most unique classes I took at Stanford was a product design elective called “Elements of Form”. Nominally helmed by my legendary advisor Rolf Faste, it was mostly run by a stylish Italian designer named Massimo, or Maximo, or something equally cool.

Each class he would dim the lights, fire up the projector, and start showing images of objects on the screen. He’d point to one of us, say “Go”, and that person would start describing the object in as much detail as possible.

Ok, it’s a green, globulous, um, shiny, um, vase thing, um…”

As soon as that person ran out of steam, he’d point to someone else and have them continue where their classmate left off.

It has 4, no wait, 5 images of centaurs on it, or maybe they’re just horses? The spout opening thing is flared on one side with a silver lip around it, and, um…”

The first few sessions were pretty painful as we struggled to turn what our eyes saw into clear, descriptive language. But as the quarter went on, we started to get the hang of it. We developed techniques for separating out parts from the whole, picked up good descriptors from each other, and generally learned the language of industrial design, something most of us had only intuited before.

I’ve thought about that class a lot in the past year as AI has upended a lot of software design practices. Now that generating visuals and code is easier than ever, the burden shifts toward expressing what you want. Most of us when using AI tools have an idea in our head of what we want, but the tools can’t access that. The bottleneck is turning those ideas into language that LLMs can work with. Developing and maintaining your vocabulary of design remains an important task for professional and aspiring designers alike.

I also think that we’ll develop tools that let you express your desires in other ways; I’m working on one myself that’s a highly visual experience. But I believe there will be lasting value in being able to turn your thoughts into the language of design.

When to say no

One of my favorite tv moments was from Richard Branson’s short-lived “Rebel Billionaire”, where he auditioned candidates to take over the presidency of the Virgin Group business.

Given his reputation as a thrill-seeker, it was no wonder that many challenges involved pushing physical limits and taking risks. But one challenge took a different approach:

It’s an example I often reference (just yesterday with my kids) for the importance of setting limits and saying no to things that others might ask of you, even leaders and people you respect. Know what you believe is important and stick to it under pressure.

Maybe attention IS really all you need?

The famous paper that kickstarted the LLM revolution found that paying simultaneous attention to everything around a word helps computers understand its meaning far better than other approaches. Where its attention was really mattered.

However, while computers thrive at processing lots of things simultaneously, we humans are notoriously bad at it. Our brains immediately throw away most of the input they receive, in order to focus on what’s most important to them.

AI tools have dramatically expanded my reach as a designer. I now have advanced business strategy and technical capabilities at my fingertips, and it’s given me the ability to build lots of new things. However I find it interesting that I still spend most of my time–my attention–on the interface; the experience; the “design”.

That, I suspect, will continue to be the case for different roles working together on products. Yes, we all have the same access to ever-expanding knowledge and capabilities. And we should all take note of the opportunities and constraints in different areas. But with limited human time and attention, we’ll each end up spending it on the things we are most interested and capable in.

To make something truly great, you’ll still need people who pay most of their attention to the business; others who pay most attention to the technical capabilities; and yes, people who pay most attention to the design.

So maybe attention–in the right places–is all you need.

Boring excellence

Giannis Antetokounpo:

Being good at something is boring. The best players, best athletes, never get bored. Every day for 15-20 years.

Or alternatively, “extremely simple and incredibly difficult“.

Finding truth in stories

The grifters and the hucksters and the influencers selling impossible things succeed because audiences reward certainty and punish doubt. They honor confidence and resist complication. A clean story about a genius who will fix everything travels faster than a difficult story about tradeoffs. The Field of Miracles stays open because people keep wanting to bury their coins there. – Om Malik

Related to themes I’m currently reading in More, Everything, Forever, by Adam Becker. Stories are powerful, and we are often all too willing to let them mislead us.

Dimensions of experience

Last week I helped with an activity in my fourth-grader’s classroom. While sitting in a tiny chair at the back of the room, my eyes scanned the rich environment around me. Drawings, sculptures, bookshelves, markers, desks and tables; every color and shape imaginable; kids excitedly talking and working.

There was so much diversity, my eyes had trouble resolving it all. I realized then that due to my work and information sources, I spend a large majority of my time looking at flat screens 18-24 inches from my face. My eyes simply are not used to changing focus or direction in order to do work anymore. This isn’t made any easier by their increasing age, of course.

I spoke with a friend later in the week who described their high school-aged son’s schoolwork as entirely screen-based now. “He lies down sideways on the couch with his laptop and doesn’t move for hours,” he said. “All of his assignments and learning materials are delivered via the screen, and he stays in one place until he’s done.”

I’m sure that’s efficient, though hearing it I felt fortunate to have gone to high school before computers became the dominant medium. I’m also lucky to have hobbies and kid activities that do give me dimensionality in the rest of my days. But when such a big part of life is spent in flat experience, it has an impact on how we think, feel, and act.

Julie Beck wrote last year in the Atlantic about “the great friendship flattening“. She noticed that when friend updates were mixed with work, entertainment, and news content, it made those friendships feel flatter and hard to distinguish:

When my phone does its little mating calls of pings and buzzes, it could be bringing me updates from people I love, or showing me alerts I never asked for from corporations hungry for my attention. When I pull it out, content and communication appear in similar forms—notifications, social-media posts, vertical video—and they blur together.

Tom Vanderlinden analyzed recent trends in cinematography and noted that modern movies tend to have a flatter, less realistic look to them. He identified perceptual realism, indexicality, haptic visuality, and cinematic qualia as choices that make the viewer feel a movie is more or less “real”.

One aspect in particular is the depth of focus used–do viewers see everything in detail across the screen, or only the central actor or item with the background blurred. This kind of hyperfocus mimics our own behavior when looking at personal devices; the world behind them is blurred, and only the interface is in focus.

Vanderlinden also quotes Laura Marks from The Skin of the Film, where she says:

Film is grasped not solely by an intellectual act but by the complex perception of the body as a whole. This view of perception implies an attitude toward [a film], not as something that must be analyzed and deciphered in order to deliver forth its meaning but as something that means in itself.

Marks describes “haptic visuality” as a way of seeing that “tends to move over the surface of its object rather than to plunge into illusionistic depth, not to distinguish form so much as to discern texture. It is more inclined to move than to focus, more inclined to graze than to gaze.”

I can’t remember the last time I “grazed” an interface on a screen; interfaces overall have been so hyper-optimized for efficiency that the very idea of it seems an anathema. And yet I feel a craving for that sort of richness of experience; that our screen lives today are poorer without it. Our minds know that an experience is more than just the optics, that real things should be engaged with all of our senses.

So what might this mean for digital designers, who focus on screen-based experiences? Are screens inherently limiting to human experience? Should we return to skeuomorphism in an attempt to engage the “haptics” of the eyes? Might we need virtual reality goggles to take our interfaces to a higher level? Should we all just go touch grass?

I believe flat, rectangular screens will continue to dominate our informational and creative lives. After all, flat books have sufficed for storytelling since Gutenberg, and my own thoughts on the challenges of face computers are well documented.

But it seems worth thinking about visual richness, realism, and dimensionality as important aspects of the human experience that we aren’t currently getting from our screens. And if screens can’t bring them to us, then we will need to more often raise our eyes from them to refocus.

Ask how,  not why

More ways to save democracy by talking about it…

Ask people to explain “how” their ideas would work, rather than asking “why” they believe in them:

Curiosity is contagious. There’s a “mirroring effect” in which the inquisitiveness is reciprocated: “Because you were curious, now they’ll want to be curious back.” If we avoid challenging people with “why” and instead ask for more information, he says, “Eventually you’re going to find that I might disagree with 80 percent of what you’re saying, but here’s the 20 percent where we see that there are connection points.” – Spencer Harrison

Recognize that people in your own “party” likely have more diverse beliefs than you might expect, and consider how you’d advise others to have a conversation with someone who holds a different view:

“One of the many problems with extreme partisanship is that you assume that your group is more homogenous than it actually is, and because of that we come rigid…This is about allowing people to be black sheep in their groups where they can think differently.” Todd Kashdan

Challenging these efforts are the ways media and recommendation algorithms polarize groups of people, to the extent they don’t even talk to each other anymore:

Not only are young people growing up very native to screens, to social media, and to these algorithmic impulses that push them into completely different worlds…[but also] COVID did a massive disruption for this younger generation that actually created for them an inability to even talk to each other in the normal ways…

When you have young [people] who not only don’t live together, but don’t consume the same media, don’t get the same algorithmic feeds…they’re not talking and therefore not converging. – Sarah Longwell

Time to touch grass!

Count on the surprises

I remembered air-raid drills when I was in elementary school, how we knelt, heads down against corridor walls with our bare hands supposedly protecting our bare necks, hoping that if nuclear war ever happened, Los Angeles would be spared. But the threat of nuclear war is gone, at least for the present, because to our surprise our main rival, the Soviet Union, dissolved itself. No matter how hard we try to foresee the future, there are always these surprises. The only safe prediction is that there always will be. – Octavia Butler

A hot take on hot takes

I was thinking today about how LLMs return strongly-worded responses quickly, and whether that counts as a “hot take.” The origin of the phrase “hot take” is about as messy as most hot takes themselves, but the current definition refers to both the speed at which an opinion was produced and the level of provocation it contains.

While LLMs do respond quickly, and exude confidence in their replies, the content is almost by definition average; they seek to find patterns, connections, and consensus across their training data. Even when you prompt them to create hot takes, those are simply extracted and synthesized from existing opinions online.

For now, the hot take may be a uniquely human capability. If you want a provocative, counterintuitive opinion about something, ask a messy, human brain.

Personally I tend to aim for more of a “simmer take”–it takes a bit longer to build up the heat, but when it’s done there’s a bit more consistency to the texture. But I still appreciate a good hot take for getting conversation going.