Episode web page: https://bit.ly/4fvszkT
Episode summary:
In this episode of Insights Unlocked, Nathan Isaacs sits down with three UserTesting voices—Lija Hogan, Amrit Bhachu, and Mike Mace—for a roundtable on the conversations they're hearing most often from enterprise leaders in the back half of 2026. Drawing on months of customer calls and industry events, the group unpacks where AI is actually changing how teams work, where the hype has gotten ahead of reality, and why human judgment keeps showing up as the differentiator no matter how capable the models get.
The conversation moves from the "next bottleneck" debate—is it code, is it customers?—to the growing skills gap facing junior researchers and designers, the fading subsidies behind "token maxing," and why AI in the loop, not human in the loop, might be the better way to think about accountability. They also dig into agentic AI and the emerging question of whether customers will engage brands directly or through their own AI "info butler," why testing the personality and relationship of an AI product matters as much as testing its accuracy, and what leaders should actually prioritize as they head into 2027.
You'll learn:
Why "what's the next bottleneck" is dividing opinion among AI's loudest voices
How shrinking model subsidies are forcing leaders to rethink where AI actually saves money
Why the panel prefers "AI in the loop" over "human in the loop"
How agentic AI and MCP could reshape whether customers deal with your brand or their own AI assistant
Why evaluating the personality and relationship of an AI product matters as much as evaluating its correctness
What UX research and design teams should prioritize in the second half of 2026
Resources & links
Lija Hogan on LinkedIn (
https://www.linkedin.com/in/lija-hogan-894769/)
Amrit Bhachu on LinkedIn (
https://www.linkedin.com/in/amritsbhachu/)
Mike Mace on LinkedIn (https://www.linkedin.com/in/mikemace/)
Nathan Isaacs on LinkedIn (https://www.linkedin.com/in/nathanisaacs/)
Learn more about Insights Unlocked: https://www.usertesting.com/podcast
