When AI can build anything, knowing what to build matters more

Building faster isn’t the same as building better

The morning workshops put that idea into practice.

Designers explored AI and design thinking with Bobby Hughes and Eli Woolery. Product teams worked through AI-enhanced product creation and faster feedback loops. Researchers examined how human imagination and artificial intelligence can work together and how research strategy needs to evolve as AI becomes embedded in the work.

The common thread: AI can accelerate the work. It can’t decide what work deserves to be done.

As attendee Jenny Choi put it afterward in a LinkedIn post, “Building is no longer the hard part. Deciding what is worth building is.”

That distinction matters.

When producing another prototype costs dramatically less, teams can test more ideas. But without customer evidence, they can also produce more of the wrong ideas faster.

The insight gap gets expensive

Baran Erkel, UserTesting’s chief strategy officer, put the shift in stark terms: As AI collapses the distance between an idea and a functioning product, the bottleneck moves. The question is no longer simply whether a team can build something. It’s whether it understands customers well enough to know what to build and whether the resulting experience actually works.

That creates what Baran called the Insight Gap: decision-making can accelerate dramatically while the supply of customer evidence remains relatively flat. And as software becomes more dynamic and agentic, traditional testing gets harder. There may no longer be a single, predictable path to design and QA. Teams have to watch real people use these experiences.

Baran also introduced ARQ, or AI Relationship Quality, which looks beyond technical performance to five dimensions of the human-AI relationship: understanding, trust, control, outcome and affinity.

Royal Caribbean: turning research into a product habit

Royal Caribbean already had sophisticated analytics and quantitative testing for its rapidly growing guest app. What it lacked was the “why” behind the behavior. A lean, two-person Digital Insights team used UserTesting to bring that qualitative voice into the product cycle—before features shipped. Studies that could take weeks to organize onboard could now return results within hours, helping the team scale research 10x in a year.

But the bigger win was cultural. Research became embedded in regular Product and Digital Analytics conversations. Eventually, product reviews began ending with teams saying, without prompting, “We should research that.” Customer evidence was no longer an extra step. It had become part of how Royal Caribbean made product decisions.

Cathay: when making becomes easy, conviction becomes scarce

Ernest Hui, head of design at Cathay, compared today’s AI moment to the arrival of photography in 1839. Painters feared the camera would make them obsolete. Instead, photography freed artists from simply reproducing reality and helped open the door to new forms of expression such as Impressionism. Ernest argued that AI could do something similar for design. When AI can draft, iterate and prototype at enormous scale, making is no longer scarce. What becomes scarce is conviction: knowing what is right for the customer, having evidence to support it and being willing to move the organization in that direction.

Cathay’s own transformation shows what that looks like. Research helped its design team move from a “black box” that produced screens to a strategic partner that shapes decisions. In one example, researchers discovered why 6 in 10 people abandoned a Cathay Holidays booking: an important option to change flights existed, but customers couldn’t find it. A relatively small redesign contributed to a 438% increase in bookings. Another research-led redesign of Cathay’s membership dashboard contributed to a 223% increase in status upgrades and a 57% increase in member revenue.

The lesson was simple: watch and talk with your customers before you build.

AI may give teams the ability to create working experiences faster than ever. But speed isn’t the same as knowing what deserves to exist.

Tools create possibility.

Customer evidence creates the conviction to choose what comes next.