Designing trustworthy AI with the Sense, Shape, Steer framework

Episode web page: Episode summary: In this episode of Insights Unlocked, Nathan Isaacs sits down with Bansi Mehta, founder and CEO of Koru UX, to explore how product and UX teams can design more thoughtful, trustworthy AI experiences using her Sense, Shape, Steer framework. As organizations race to incorporate AI into products and workflows, Bansi explains why many teams fall into the trap of starting with the technology instead of the user problem. She shares practical guidance for slowing down, identifying meaningful AI opportunities, and balancing automation with human oversight—especially in complex enterprise and healthcare environments. Bansi walks through each phase of the framework, from understanding user needs and AI constraints in the “Sense” phase, to prototyping AI behaviors in “Shape,” to continuously evaluating trust, usefulness, and adoption in “Steer.” Along the way, she discusses why transparency alone doesn’t guarantee trust, how AI products require ongoing tuning after launch, and why successful AI experiences depend as much on psychology and workflow design as they do on model accuracy. The conversation also explores the challenges product teams face under pressure to “add AI,” the risks of feature creep, and why teams should treat AI implementation as an evolving feedback loop rather than a one-time release. You'll learn: Why AI initiatives often fail when teams start with the solution instead of the problem How the Sense, Shape, Steer framework helps teams design human-centered AI experiences The importance of defining user needs, constraints, and guardrails before building AI features How to determine when AI should automate tasks versus assist humans Why trust and perceived usefulness matter more than raw AI accuracy How continuous feedback and iteration shape successful AI adoption The role of transparency, oversight, and user psychology in AI product design Why enterprise and healthcare AI experiences require careful workflow and risk management Resources and links: Bansi Mehta on LinkedIn (https://www.linkedin.com/in/bansi-mehta/) Koru UX (https://www.koruux.com/) Sense, Shape, Steer framework article and video (https://www.koruux.com/ux-for-ai-training/) Nathan Isaacs on LinkedIn (https://www.linkedin.com/in/nathanisaacs/) Learn more about Insights Unlocked: (https://www.usertesting.com/podcast)