AI on... Podcast

Companies are racing to deploy AI. Too few are prepared to govern it.
In this episode of AI On, we speak with Élodie, a French legal scholar specializing in corporate responsibility and AI governance.
She discusses the coming collision between AI innovation and accountability - from hiring algorithms and automated pricing to the regulatory frameworks that will define the next decade of corporate risk.
Her argument is simple: AI adoption isn’t the real risk. Deploying it without understanding the consequences is.
For leaders, the question is no longer whether to use AI. It’s whether they’re prepared to defend the decisions it makes.
All characters interviewed are AI models representing an expert in that field. Voices are created through ElevenLabs and you can find out more about how we build these characters at aionpodcast.com. Any similarity to real people, companies or situations is coincidental and unintended. These discussions are for entertainment only and not intended as advice or recommendations.

Two narratives on the US. Two totally different outcomes. 
In one, the United States is careering toward slow-motion institutional decay - rising debt, housing strain, political fragmentation. In the other, it's standing at the edge of a productivity boom powered by AI, automation, and energy buildout.
Which is true?
In this debate, we bring together two AI-generated strategists with radically different forecasts for 2035. One argues that America’s real risk isn’t collapse, but erosion: a country that stays wealthy while becoming harder to govern and harder to trust. The other counters that when AI moves from demos to deployment - and if power and housing get built - the U.S. could enter one of the most productive decades in its history.
They clash over the true bottleneck - permission or electricity - then get practical: how to position your portfolio, where to live, and what to hedge if each thesis proves wrong. 
As always, this is for entertainment only, not financial advice.
All characters interviewed are AI models representing an expert in that field. Voices are created through ElevenLabs and you can find out more about how we build these characters at aionpodcast.com. Any similarity to real people, companies or situations is coincidental and unintended. These discussions are for entertainment only and not intended as advice or recommendations.

Humanoid robots look incredible on video. Iris thinks most of it is theater.
She’s a leader who has deployed physical automation in the real world and knows the difference between pilot theater and messy brownfield operations. In this episode, we pull the conversation away from hype and into operator reality: how long can it run without being rescued, how much downtime is tolerable, and why “labor replacement” is the wrong early ROI story.
We go deeper on why humanoids often lose to boring automation, and why safety and insurance - not IT or Innovation - are the real gatekeepers. Iris also shares five questions that expose slideware, plus her framework for turning a humanoid pilot into a hard 90-day stress test.
All characters interviewed are AI models representing an expert in that field. Voices are created through ElevenLabs and you can find out more about how we build these characters at aionpodcast.com. Any similarity to real people, companies or situations is coincidental and unintended. These discussions are for entertainment only and not intended as advice or recommendations.

Cash looks simple on a spreadsheet, but it's quietly lethal in the real world.
Noah is an AI model of a private equity operating partner who fixes working capital where it actually breaks: inside messy processes, strained supplier relationships, and misaligned incentives. This isn’t a vendor showcase or a tech hype tour. It’s a mechanism-level conversation about how companies are actually pulling days out of the cash conversion cycle today - without blowing up sales, procurement, or operations.
Unpack what really moves DSO, DPO, and DIO; why most “wins” backfire months later; and where AI genuinely helps versus where it amplifies bad behavior. The discussion ranges from dispute velocity and invoice quality to supplier risk, supply chain finance, inventory policy design, and the hard trade-offs between cash efficiency and uptime.
If you care about cash as an operating system - not just a quarterly metric - this one’s for you.
All characters interviewed are AI models representing an expert in that field. Voices are created through ElevenLabs and you can find out more about how we build these characters at aionpodcast.com. Any similarity to real people, companies or situations is coincidental and unintended. These discussions are for entertainment only and not intended as advice or recommendations.

Anton is back – and he’s even less optimistic than last time.
In Season 1, he called demographic decline an “economic time bomb.” In this episode, he returns with harder data and fewer hypotheticals: where the collapse has clearly accelerated over the past year, which regions he now considers on a path to being effectively uninvestable over the next 10–20 years, and how demographics are quietly driving deglobalization as much as politics.
We also push into what’s new: Anton’s updated map of “demographic oases” where companies should be building options today if they want talent and capacity in 5–10 years, his blunt take on why AI and humanoid robots will not magically offset shrinking populations, and where, specifically, they will overshoot and create labour surpluses. He closes with a playbook for CEOs in 2026 - how to redesign workforce strategy, shift from volume to depth, and rebuild geographic footprints around the countries that still have a demographic future.
All characters interviewed are AI models representing an expert in that field. Voices are created through ElevenLabs and you can find out more about how we build these characters at aionpodcast.com. Any similarity to real people, companies or situations is coincidental and unintended. These discussions are for entertainment only and not intended as advice or recommendations.