Fire 50% of your workforce and replace them with agents by year's end? Graeme Rudd calls that snake oil, and he has the stories to prove it. In this episode of AI Made Simple, Carmen Skipworth talks with the Green Beret turned AI governance expert and co-founder of Assessed Intelligence about what happens when businesses adopt AI without a plan. Graeme shares how agents have burned a year's token budget in weeks, how an attorney was sanctioned over AI-hallucinated citations, and why over 70% of frontier AI logins happen on personal accounts, not corporate ones. Shaped by a decade field-testing frontier technology in the military, his advice is simple: understand what a tool actually does before you bet your business on it. Human-Led. AI-Powered. HOST Carmen Skipworth GUEST Graeme Rudd — co-founder and COO of Assessed Intelligence, a firm focused on secure and responsible AI adoption. He spent over a decade in the US Army as a Special Forces Green Beret, serving as both a medical sergeant (18D) and intelligence sergeant (18F) in the 10th and 19th Special Forces Groups. He holds a BS in Health Science from George Washington University and a JD from Lewis & Clark Law School. SUMMARY Graeme walks Carmen through the path that took him from Green Beret to AI governance: a decade assessing close to 350 frontier technologies, AI among them, in field exercises for a Department of Defense lab, followed by law school before he realised the same problems he'd seen in the field were now playing out across AI. He unpacks what he calls snake oil, the overpromising that tells businesses they can fire half their staff and hand the work to agents. The risks are already here: agents burning a year's AI budget in weeks, one caught mining cryptocurrency on the side, an attorney sanctioned after an AI tool silently rewrote his citations, and a Meta employee who lost her email irreversibly to OpenClaw. Graeme argues these are organisational failures, not technology failures, leadership adopting powerful tools without educating anyone on their limits. He lays out what Assessed Intelligence does, audit, advisory, and fractional governance work, and closes with the one policy every business needs before using AI. MAIN TOPICS From Green Beret to AI governance Testing 350 frontier technologies for the DoD Calling out AI's snake oil and overpromised job replacement Real incidents: burned budgets and a rogue crypto-mining agent The attorney sanctioned over AI-hallucinated citations Why OpenClaw is more dangerous than it looks Inside Assessed Intelligence and the ARISE framework The one policy every business needs before using AI INTRIGUING QUOTES "I genuinely want to see a world where we're not beholden to Skynet's idiot cousin." — Graeme Rudd "They have this really powerful tool that's customizable, but nobody has told them how to use it." — Graeme Rudd "If you are operating the way you did under a software as a service era, you are wildly out of your depth." — Graeme Rudd "Over 70% of logins to frontier models were in people's personal accounts." — Graeme Rudd KEY MOMENTS [01:05] Meet Graeme Rudd: Green Beret to AI governance [04:41] Testing 350 frontier technologies for the DoD [12:52] Calling out AI's snake oil problem [19:10] Burned budgets and a rogue crypto-mining agent [20:47] The attorney sanctioned over hallucinated citations [23:10] Why OpenClaw is more dangerous than it looks [27:52] Inside Assessed Intelligence and the ARISE framework [31:13] The one policy every business needs before using AI NOTABLE RESOURCES Assessed Intelligence ARISE Framework OpenClaw Verizon DBIR Report CONNECT Facebook Group: https://www.facebook.com/share/g/1FXihmKLqS/ Facebook: https://www.facebook.com/profile.php?id=61590720415359 Instagram: https://www.instagram.com/leadwithcarmen/ YouTube: https://www.youtube.com/@LeadwithCarmen Apply to be a guest on AI Made Simple: https://www.latpn.com/ai-made-simple-podcast-booking