Photo: Barak Sharma / LtoR: Patricia Roller of consumer research firm Vidlet; Paul Wood of Alice; David Flynt of Lucid Motors; Frank Bach, formerly of DoorDash and now with Anthropic; and Nicole Kwan Leighton, Alice
Today, product design teams are facing increased pressure to move faster without sacrificing confidence in decision-making. Even projects that begin with extensive upfront research inevitably generate new questions once design work is underway. But here’s the problem: Traditional research methods often can’t keep pace with modern product cycles, forcing teams to choose between speed and certainty.
With that in mind, Paul Woods, the founder/chief creative officer at independent design agency Alice (and a Graphic Design USA 2026 Person to Watch), convened a panel together with research firm Vidlet, of some top minds in the design industry to discuss how brands are incorporating real user signals into the design process, even while projects are already in motion.
.
.
In addition to Woods, the panel, which met in San Francisco this month during the recent Figma Config event, included product design experts Frank Bach, formerly of DoorDash and now with Anthropic; David Flynt of Lucid Motors; and Patricia Roller of consumer research firm Vidlet.
.
.
Here’s an excerpt from that conversation.
.
Question: Looking at research within the product cycle, how do great teams build user insight into the work as it happens?
David Flynt: Insight has to fit the lifecycle phase. At Lucid, early in the lifecycle, hardware is not available, so we have to create high-fidelity simulations to evaluate concepts in the driving environment. The goal there is to understand how potential solutions map to needs but also can succeed in the target environment. In one study, we had two interaction models under discussion; model A was the paradigm, and B was unique. The designer’s instinct was to use the unique version. When we tested both, users told us they preferred prototype A, but the behavioral and interaction metrics from the study made it clear B was better all around. That’s where designers earn their pay: turning a gut reaction into something you can actually act on.
Patricia Roller: Teams need to know where they are in the cycle. The most common mistake I see is confusing the empathy phase with the testing phase; people think they’re already testing something real when they’re not. It comes down to asking the right question at the right time and being willing to keep asking ‘why.’
Frank Bach: Research has always been about de-risking design, but the approach changes with context. At a large platform you can have almost too much data, while at an early-stage startup you begin with nothing. In one role, thousands of app-store reviews surfaced clear themes without me ever hopping on a call; in another, feedback flowed constantly from both drivers and customers. You have to match the method to what you’re working on.
Paul Woods: In my experience, user insight should be built in at all the key intersections where there’s a decision point in your process, before you move on to the next phase. The trick is that you need to already have a couple of working hypotheses you can put in front of people to elicit a response; overly theoretical questions don’t work. There’s no substitute for putting a prototype, no matter how rudimentary, in front of a user and getting their reaction. As the old saying goes, products or features don’t fail because of a bug. They fail because nobody needed them in the first place.
.
.
Q. If that’s the case, then what changes when insight arrives in a day instead of a month? Talk about leveraging new technologies for qualitative speed.
Frank Bach: AI is a double-edged sword. It makes research far more accessible, because you can point it at many sources at once, but if you don’t understand the context, you can lose the very thing you were looking for. For me, the real value is being able to spin up a dozen prototypes and test them quickly.
David Flynt: Teams rarely struggle to collect telemetry; making it land is the challenge. The tools we use now dramatically reduce the time between inquiry and insight. That allows us to rapidly identify design opportunities, craft solutions, and test to see if our ideas are on target.
Patricia Roller: As AI speeds up collection, human interpretation is where the value concentrates: knowing the right question to ask, and continuing to ask “why.”
Paul Woods: Faster insight makes creative decisions easier, but only if the question is good and the timing is right. As Dave mentioned, users usually can’t verbalize their problem; the pseudo-Ford line, “If I had asked people what they wanted, they would have said faster horses,” comes to mind. So as designers we have to know what to take literally and what to interpret. The customer doesn’t know what they want until they see it, so you have to give them something to react to. More insight isn’t the goal; better insight is, and that comes from the right question with a prototype as the stimulus.
.
Q. So, what makes user insight actually change a decision? What are the user signals that drive stakeholder alignment?
Patricia Roller: Video is the most effective tool I’ve found for aligning stakeholders, especially in board meetings and workshops, because it brings the customer into the room. C-level audiences want to be entertained too, so we now pair every insight with a real user on video, and it lands. Honestly, the biggest weakness in our industry isn’t the research; it’s the ability to sell it.
David Flynt: We adopted voice and video for the same reason. Our early reports had plenty of data but needed more sizzle to connect. The insight that actually moves a decision is the data that’s been internalized, so the real issue was getting that insight into a form that could land with a leadership that tends to think they are the user. Seeing a video of a real user speaking is a lot harder to ignore. First-person experience also matters, so getting insight out of reports and into prototypes is crucial. When one of our executives who doubted a design insight finally tried a prototype, they realized they were actually one of the users we were designing for.
Paul Woods: Defining product nomenclature is a great example. Time and time again, large organizations approach naming conventions from an inside-out perspective rather than a user perspective. But as soon as those names are put in front of users, we hear how people actually describe their journey and pain points in their own words, which is usually quite different.
Frank Bach: Success can breed resistance. I’ve worked with founders who were skeptical of research precisely because the company’s early wins had gone so well. The real hurdle is making qualitative insight credible in a metrics-driven culture.
.
Q. Lastly, what does your industry understand about user insight that others could learn from?
David Flynt: Hardware products can take years to ship, so the discipline is knowing where you’re comfortable living with uncertainty for differentiation and where you genuinely need certainty around parity experiences. The future state in which your product must succeed will be impacted by regulation, environmental conditions, new technology, and competitive experiences. That timing shapes how you deal with every piece of feedback you get.
Frank Bach: How much certainty you need depends on the decision. Different categories teach different lessons, and the environment itself biases feedback. I once ran testing inside a brand’s flagship headquarters, and the space was so immersive that it skewed how people responded; you’re never getting a neutral read in a setting like that. Cultural differences matter too.
Patricia Roller: Organizations fool themselves into feeling certain, and the better established the company, the harder it is to change course. Startups tend to be more honest about how little they actually know.
Paul Woods: There’s always uncertainty; it’s just part of building products. What matters is putting the metaphorical pen to paper by generating scenarios and prototypes you can put in front of people. That immediately turns something uncertain and theoretical into something real.
