UX Institute Podcast: Why 85% of AI Projects Fail - And Why Most Teams Are Still Getting AI Wrong

From the Description:

"If you work in UX, product, or AI and feel like the ground is shifting under your feet, this is the kind of conversation that helps cut through the noise. Greg’s argument is simple but urgent: most AI projects are not failing because of the models. They are failing because teams are still choosing the wrong problems, using the wrong data, and applying outdated product habits to a completely different kind of technology."

"This episode is worth listening to because it reframes where real value now sits for designers and product people. If your role has drifted into handoffs, surface-level screens, or feature packaging, AI will expose that quickly. But if you can frame the right use case, understand customer pain, evaluate risk, and test fast with something real, your role becomes more valuable, not less. Greg brings sharp, practical thinking to what AI teams are still getting wrong and what UX and PMs need to do now to stay relevant."

In this episode, you’ll learn:

00:00 - The 85% Failure Rate: Why Enterprise AI Projects Keep Collapsing

05:21 - The Hadoop Syndrome: When Boards Demand AI Without a Strategy

07:32 - Robot Monkey Work: Have Designers Made Themselves Disposable?

12:56 - Horseman 1: Torpedoing Your AI Project with the Wrong Use Case

14:18 - Horseman 2: The Data Delusion and Building a Bullshit Generator

20:52 - Horseman 3: The Value Matrix and the Hidden Cost of AI Accuracy

22:28 - Horseman 4: Snowball Sprints and Killing the Traditional UX Handoff

26:44 - The Disruption of Product Management: Why AI Will Replace Bad PMs

32:24 - Becoming an AI MacGyver: Why Vibe Coding Beats Figma Prototypes

36:20 - The Six-Week Proof of Concept: Starting with Knowledge Management

Listen on Apple Podcast: