What if 40% of your workforce is quietly correcting AI hallucination every single month and the productivity gain you booked never actually arrived?

Edosa Odaro and Lindley Gooden sit down on Episode 8 of The AI Values Podcast with David Edam a data and AI transformation leader who has spent two decades inside some of the most operationally complex organisations on the planet — for an uncomfortable conversation about AI governance, AI risk management and the hidden work of generative AI in the enterprise. This is not the demo-day story. It is what happens after the hype dies down, the boardroom goes quiet, and someone has to make AI actually deliver value in regulated industries where mistakes have real-world consequences.

THIS WEEK: A GUEST EPISODE
Our guest is David Edem, a data, cloud and AI transformation leader who has worked across global oil & gas, energy infrastructure and global SaaS building production AI inside multi-billion-pound asset environments where decommissioning, safety, and engineering integrity are not theoretical. David brings the practitioner's view of what changes when AI meets operational complexity, and why "the last 10%" is the part most leaders quietly misjudge.
◼ Guest LinkedIn: https://www.linkedin.com/in/david-edem/

WHAT IS COVERED:
► Why AI doesn't replace work it relocates it, and the hidden work of generative AI is now showing up in every team's calendar
► The Microsoft + Workday "workslop" stat: 40% of workers correcting AI-generated errors every month, and what it really means for AI productivity in 2026
► Inside oil & gas: how decades of engineering drawings, seismic data and well trajectories became a one-year AI project instead of a several-year, multi-million-pound rebuild
► Innovators, sceptics and the delayers nobody talks about: the third group in every boardroom that quietly kills AI decisions
► Why you should never walk into a boardroom selling AI — you walk in selling the outcome (boardroom storytelling for AI adoption)
► "AI wants you to be happy. It gives you good results and it makes stuff up." — David on AI hallucination, confidence, and why the last 10% of AI requires humans
► Efficiency vs trust: the tension every C-suite is missing, and why human in the loop is not a soft option but the only credible posture
► How the barriers to entry collapsed: the customer success manager with no technical background who became a global AI product leader in 18 months
► Why AI doesn't only replace work it creates work that wouldn't have existed (David's own counter-example to the displacement narrative)
► A practical framing for responsible AI in regulated industries: data quality, human oversight, and the AI value extraction question every board should ask

⏰ EPISODE TIMESTAMPS:
00:00 — Cold open: "AI wants you to be happy" — and it makes stuff up
01:30 — Microsoft's 40%: workers are fixing AI every month (the workslop problem)
04:50 — Meet David Edem: AI inside operationally complex industries
07:10 — Oil & gas, decommissioning, and AI as a shortcut to value
11:40 — Innovators, sceptics — and the delayers nobody talks about
14:00 — Don't sell the AI, sell the outcome — the storytelling gap
16:20 — Hallucination and confidence: why the last 10% needs humans
21:10 — Barriers collapse: from no-code customer success to AI product leader
26:00 — AI doesn't replace work — it creates work that didn't exist
28:20 — Debrief: storytelling, delayers, and what's really at stake