Why Rushing AI Products Without User Research Is a Bad Idea

Eric Blumberg started out as an electrical engineer and quickly found that electrons weren’t his thing, but the experiences they powered were. In undergrad he discovered human factors, worked in a research lab, and went on to earn a PhD in the field. At Apple he worked on Face ID and multi-touch gestures for the iPad, then moved into fintech, SaaS and B2B experiences, most recently as VP of User Experience Research at Fidelity. He’s also involved in his local UXPA conference in North Carolina, which brings together more than 300 people from across the region each year.

His view of AI and research comes down to one distinction. AI is great at knowledge and execution: storing data and delivering it on demand. What it lacks, and what research specializes in, is the wisdom and judgment to know when a piece of data applies and which tradeoff to make. Teams moving fast without that judgment build polished solutions that miss the actual problem, and businesses lose money on features nobody uses. AI is already a clear win for analysis, like pulling themes from raw data with a researcher still in the loop, and for the operational overhead of documentation and project tracking. It can’t replace the early conversations about which problem the business should solve, or the human work of communicating insights to stakeholders. He points to a recent study finding that AI moderators talk far more than human moderators, collect less data, and engage some people less well, which makes the data less representative.

When product teams say “skip research, we’ll ship to production and get feedback,” his answer is to push back by setting the right context: framing problems in terms of value to the customer and the business, and building guardrails into how the organization works. He encourages researchers to use AI to speak the language of design and product partners without pretending to be them, and to ask what they’d do with the 20 percent of their time AI frees up. For people entering the field, research is still valuable, though it will look different in five years. As technology makes polished solutions a commodity, the differentiator is removing friction so people reach their goals faster. The best organizations for researchers, he says, treat UX as specialized expertise rather than a function to bypass.