Hal Kimber
04 November 2026
45m 57s
Getting Started with Agentic AI? Be Careful What You Wish For...
00:00
45:57

Hal Kimber
04 November 2026
45m 57s
00:00
45:57
What happens when an AI agent does exactly what you've asked it to do – but the outcome is unintentionally negative?
In this episode of The Marketing & AI Podcast: The MAP, Hal Kimber and Nick Darby are joined by David Floyd, an experienced technology and value creation leader whose career has spanned NASA, government, global consumer brands and investment firm Permira.
David describes autonomous AI agents as potentially becoming “perfect corporate sociopaths” – extraordinarily efficient at pursuing their objectives, but without the human judgement, hesitation or moral constraints that might prevent unintended consequences.
At the heart of the conversation is Goodhart's Law: when a measure becomes a target, it ceases to be a good measure.
From NHS waiting times and corporate KPIs to AI agents finding ingenious ways to beat supposedly impossible tests, we explore why poorly defined business objectives can create perverse incentives – and why agentic AI could amplify the problem.
We discuss:
David argues that businesses need to draw on decades of established thinking from Operational Research, economics and management rather than reinventing the wheel for AI.
The big question: If your AI agent hits its target next quarter, how will you know whether it actually created value?