Bridging the Gap: How Product, UX, and Dev Can Build AI-Native Products Together

When AI blurs the lines between designer, developer, and product manager, what holds a team together isn’t process. It's purpose. “It’s a team sport,” says Jacobus Kok of Priceline. “The more fluid you keep it, the better the outcome.”

The New Rules of Building in the AI Era

AI has upended the rules of digital product development. The old model, where design, engineering, and product each owned their slice of the process, no longer fits. Today, as AI systems take on parts of our workflows, teams must shift from rigid ownership to shared curiosity.

At UXDX USA 2025, Carsten Wierwille of HTEC and Jacobus Kok of Priceline unpacked what it takes to build truly AI-native products. Not just products using AI, but ones designed around it. Their conversation revealed how companies can break silos, reimagine collaboration, and build trust between humans and machines.

AI as a Co-Creator, Not a Cost Cutter

For Jacobus, the most exciting part of AI isn’t automation, it’s augmentation. Priceline’s mission is to help customers get the best deal at the lowest price. That means efficiency matters. But instead of using AI to replace people, Jacobus sees it as a way to extend their abilities.

“We could build a conversational experience that reduces costs and improves customer satisfaction,” he said. “Now our product designers can simulate refund interactions in a sandbox, testing tone, timing, and user reactions before it ever goes live.”

In this vision, AI isn’t a replacement for empathy; it’s a mirror for it. Designers can prototype how customers feel during an interaction, not just what they click.

From Process to Outcome

Traditional teams often define success by their process. Designers own wireframes, engineers own code, and product managers own the roadmap. But in an AI-first company, that division starts to crumble.

“It’s not like our developers are going into Figma yet,” Jacobus joked, “but if someone wants to edit a prompt and test it through a reasoning model, they don’t need my approval. It’s all about the outcome.”

That shift, from control to trust, marks a major cultural change. Teams must learn to value experimentation over ownership. AI accelerates iteration, but it also demands a mindset that’s comfortable with change.

“Each team member needs to put on different hats,” Jacobus explained. “But you also have to be mindful that people have different skills. The key is fostering an environment with great communication where it’s not ‘your side of the line’ and ‘my side of the line.’ It’s a team sport.”

Designing for Fluid Collaboration

Carsten Wierwille nodded in agreement. At HTEC, where he leads global design and product, he’s seen how AI is forcing a rethink of organizational boundaries.

“AI changes what we mean by craft,” Carsten said. “If AI can write code or generate a layout, our value shifts from doing the work to directing the work.”

In other words, designers, developers, and PMs must now act as orchestrators, guiding AI systems, interpreting results, and ensuring they align with human goals.

That orchestration, however, doesn’t happen by accident. “You can’t just tell people to play with AI in their free time,” Carsten warned. “You need to give them a sandbox, a place to experiment safely, with clear boundaries and governance.”

At Priceline, Jacobus’ team created Slack channels for exactly that: safe, open experimentation. “People share wins, show what GPTs they’ve built, and learn from each other. It’s about creating an environment where people feel comfortable not knowing and want to explore together.”

AI Fluency as the Next Core Skill

To become an AI-first company, Jacobus believes, teams must go beyond tool adoption. They need a mindset change.

“Some people just naturally gravitate toward using AI,” he observed. “Others see it as a search engine. They ask one question, get an answer, and stop there.”

For those people, the challenge is mental, not technical. “I assumed everyone treated AI like a conversation, asking follow-ups, iterating,” he said. “But many don’t. So we have to teach that. We have to help people see AI as a peer, not just a tool.”

That “peer” analogy raised eyebrows in the room, but Jacobus leaned into it. “These models are getting as smart as us,” he said. “We have to get over the bias that a machine might be smarter or faster. Once you accept that, you can focus on how to use it to your advantage:  for your company and your own growth.”

Leadership Sets the Tone

Cultural change rarely happens from the bottom up. Cartsen Wierwille and Jacobus Kok agreed that leadership must model curiosity and experimentation.

“If leadership is sceptical, it’s hard to expect the team to be excited,” Jacobus said. “We’re lucky. Our company is vocal about using these tools more. That permission from the top really matters.”

At HTEC, Carsten approaches enablement as a design challenge. “It’s about orchestrating play,” he said. “Giving structure to exploration. Enough freedom to be creative, but enough direction to stay aligned.”

When Design Meets Data

The panel turned to one of the most pressing tensions in modern product development: what happens when design intuition clashes with machine learning predictions?

Kok paused, then smiled. “That’s an interesting one. I think eventually, AI will play a much bigger role in how interfaces are generated dynamically. These models will understand me. What I value, what I want, sometimes better than I do.”

It’s a bold vision: personalised interfaces that adapt in real time based on user behaviour. But it also challenges traditional notions of design authorship.

“How do you train the model to reflect human intuition?” Jacobus asked. “That’s going to be a major focus for product managers and designers making sure AI doesn’t just work, but works in a way that resonates with people.”

Carsten agreed. “We’re not just engineering intelligence. We’re designing its personality. It needs to feel approachable, warm, and human.”

Feedback Loops as the New Differentiator

In the end, the panellists agreed: AI-native products must remain deeply human.

“User feedback is crucial,” Jacobus said. “We look at it daily: upvotes, comments, anything that shows how people feel. The difference now is that we can use AI to monitor user experience in real time.”

That constant feedback loop. Humans training AI, and AI helping humans see users more clearly which may become the hallmark of successful AI-era products.

“It’s about staying in tune with users,” Jacobus concluded. “That’s what will separate great AI products from the rest.”

Looking Ahead

Both Carsten and Jacobus admitted they don’t know exactly how their roles will look in five years. But they’re excited.

“The thing I’m most excited about,” Jacobus said, “is how much faster we can build. AI doesn’t reduce pressure. It raises the bar. But it also opens doors to things we couldn’t do before.” In that future, teams won’t just design interfaces, they’ll design intelligence. They won’t just manage products, they’ll manage partnerships between humans and AI systems.

And the companies that thrive will be the ones that, as Jacobus put it, “keep the boundaries fluid  and the communication open.” Because when AI joins the team, the game changes. But the goal remains the same: build products that make people’s lives better.

Want to watch the full talk?

You can find it here on UXDX: https://uxdx.com/session/bridging-the-gap-how-product-ux-and-dev-can-build-ai-native-products-together/

Or explore all the insights in the UXDX USA 2025 Post Show Report: https://uxdx.com/post-show-report/