Research and voice-of-customer teams are being consolidated, absorbed, or cut. An AI tool usually gets named as the reason, and a CPO usually signs off. unitQ asked research professionals directly what is actually happening. 42% said AI has slowed hiring or left roles unfilled, and the work automated first is the execution work that trains people into judgment. The need did not shrink with the headcount: across 83 million app reviews, the average quality score is still 60 out of 100. When the function that owned customer understanding is one person or nobody, the responsibility does not disappear. It becomes unassigned. Christian Wiklund, co-founder and CEO of unitQ, will argue that the fix is not rebuilding the team you just cut. Product teams already build with coding agents, and those agents can read customer signal themselves once it is exposed as something they query instead of a dashboard someone is supposed to open. An agent that asks what is breaking in checkout before it writes the patch, ranks the work by what that friction costs, and validates the fix against real users the same week, with no human relaying anything in between. Christian will walk through what has to be true for that loop to close, and the three places it still breaks.