Topics
August 19, 2026
August 19, 2026
Speed has taken the front seat as AI rapidly transforms how we build products. But as insight generation becomes more distributed—and cross-functional teams spin up new tools faster than ever—how do we make sure the voice of the customer doesn't get lost in the noise?
We recently sat down with research executives from Expedia, Cisco Networking, Dscout, and AWS to uncover how they’re approaching these new challenges.
The consensus? AI isn't just a tool for speeding up research; it's a catalyst for raising our ambition. Here’s how top leaders are leaning into the disruption, keeping customer-centricity intact, and ensuring research leads the way.
Contributors include:
Researchers can no longer be the sole gatekeepers of quality as insights scale. The focus has shifted from guarding the process to governing the democratization of insights without becoming a bottleneck. Here's how teams are automating their rigor.
At Dscout, the team previously relied on a human-led program called "PWDR (people who do research) Donut" to teach designers and engineers research best practices.
While effective, it became too cumbersome in a high-speed environment. To adapt, they built a custom Gem that automatically reviews study plans against their internal guidelines, eventually evolving into their “Draft Your Study”feature.
Reader tip: Stop relying on manual reviews for every stakeholder study. Try using custom AI agents to bake your team's unique research standards and context directly into the templates and tools your stakeholders use to self-serve.
“We're past the democratization conversation a bit...That process just became too cumbersome. What we've shifted towards is using AI to shepherd best practices.” — Taylor Klassman, Director of Research at Dscout
Opening up a research repository to the entire company comes with the risk of stakeholders misinterpreting data. At Cisco Networking, the research team mitigates this by designing the exact structure of how AI agents return insights from their repository. They create strict frameworks that mandate how data is contextualized.
Reader tip: Don't just open your data vault and walk away. Design strict "contracts" for your AI agents so that every generated insight automatically leads with a direct answer, separates different types of evidence, and makes sure caveats are clear.
“The contract is designed to prevent overclaiming by forcing answers to lead with a direct answer, separate evidence types, expose caveats. There's a lot of rigor that's gone into the design of this framework.” — Mary Piontkowski, VP of Product Design at Cisco Networking
How do we know if customer centricity is actually landing when systems (rather than researchers) are delivering the insights? Research teams can cement trust across the organization by integrating research earlier in the product lifecycle and actively participating in the evaluation of AI outputs.
Trust begins with the quality of the AI tools themselves. Because researchers are subject matter experts inherently driven by quality, they’re uniquely positioned to serve as the critical human-in-the-loop evaluating AI outputs.
Reader tip: Take ownership of the quality of your company's AI outputs. Step into the role of evaluator for internal AI tools to ensure they meet a high threshold of accuracy and human understanding before they’re deployed more widely.
At Expedia Group, the research team avoids being an afterthought by injecting their insights into the process long before product requirements documents (PRDs) are drafted. This proactive involvement ensures that the foundational strategy is sound.
Reader tip: Don't wait for the PRD. Work cross-functionally with product and engineering from day one to build a "project context brain". This will help ensure that all the building and measuring that follow are grounded in verified customer realities.
“It's making sure that we can get research much more upper funnel, and influencing the product requirement docs and business requirements earlier in the funnel, so we can measure it throughout the entirety of build to launch.” — Rachel Been, SVP of Design at Expedia Group
Researchers can’t physically be in every meeting room. Rather than trying to helm every study, research leaders are evolving into strategic guides who curate and interpret the massive influx of data flowing through the organization.
Reader tip: Shift your team's mindset from “gatherers of insights” to “guides of decision-making.” Focus on synthesizing the noise and creating an ongoing call-and-response loop with stakeholders as new data arrives.
“The researcher can really guide the way that information flows through, and then serve as that sense-making layer who is there to help take in the copious amount of data...and help the stakeholders understand what makes the most sense, or what to use to guide decision-making.” — Alyssa White, Head of Research at Expedia Group
The true power of AI in research isn't just responding to product requests faster—it's freeing up capacity to lead on customer understanding. Researchers are using their reclaimed time to connect disparate data sources, establish foundational truths, and embrace creative exploration.
While technology changes rapidly, human psychology has largely stayed the same. Establishing a shared foundation of human behavior ensures that when cross-functional teams use AI to build and prototype rapidly, they’re doing so with a unified understanding of user limitations and needs.
Reader tip: Anchor your practice in behavioral science. When everyone builds off a shared language of cognitive limitations and human truths, rapid prototyping remains closely tied to reality.
"If we were to try and do research in a very traditional way to keep up with the iterative process as everyone is building and prototyping, it wouldn’t be possible...Rather, if we take a step back and focus on building a really strong shared foundation with our stakeholders that says, 'Here is the person, here are their limitations'... then everyone is building and prototyping with that shared language." — Alyssa White, Head of Research at Expedia Group
At AWS, the research team recognized that data is often siloed across different business units. By creating a unified vocabulary across multiple data marts, they shifted from analyzing isolated, anticipated outcomes to evaluating millions of simultaneous, real-time user behaviors.
Reader tip: Actively seek out partnerships with analytics and business intelligence teams. Do the messy work of normalizing terminology across your company's data streams so you can deliver impactful, real-time storytelling that drives executive action.
“A big program that we put a lot of investment and energy into was creating a controlled vocabulary across six or seven major data marts across the company, so we could start to bring together real-time insights from actual behavior that's happening in the product...You become a researcher of millions of simultaneous behaviors, versus someone who is anticipating a potential outcome." — Matthew Menz, former VP of CX at AWS
With automated systems handling the routine, researchers have an unprecedented opportunity to lean into R&D and innovation. By using tools like AI and data synthesis, researchers can actively prototype solutions—like Expedia Group's team vibe-coding an alien abduction product—proving that research can directly drive the product portfolio.
Reader tip: Use your newfound bandwidth to take risks. Experiment with new methodologies like vibe-coding and co-creation. You have the data that points to the future. Don't be afraid to build the first draft of it yourself.
"Find your creativity, try things out, take some risks, be bold, get into the vibe coding, because you know what we should be building. Build some of those options for teams to consider and be curious about the boundaries." — Mary Piontkowski, VP of Product Design at Cisco Networking
AI has turned almost everyone into builders, but it doesn’t mean researchers are any less important. If anything, the knowledge and experience researchers have is more important than ever—it’s deeply needed to maintain guardrails and usher in a strategic approach to this new evolution of work.
By making sure that standards remain high and humans are the foundation, customers can remain at the center of any new product building.
