Spotlight: Automating A Bad Process With AI Is Still A Bad Process

Episode notes

Companies are making the exact same mistake with AI that they made with cloud. FICO's CIO Mike Trkay breaks down why 95% of companies are failing at AI alignment, why "automating a bad process faster" is the #1 trap, and why regulated industries are already abandoning LLMs in favor of focused language models. Key takeaways: • Only 5% of AI pilots make it to production - and MIT's research backs it up • The lift-and-shift parallel: cloud costs went up for the same reason AI ROI is missing • Why the LLM-to-focused-language-model shift mirrors cloud-native vs lift-and-shift • What "AI native" actually means (and why chatbots aren't it) Chapters 00:00 The 5% Alignment Problem 01:09 Only 7% Even Measure If Their AI Works 02:15 "You're Just Doing a Bad Process Faster" 06:29 LLM Repatriation and the Rise of Focused Language Models - This episode of IT Visionaries is

• The Safest First Use of AI in Networking75 / 100
• The AI Saturation Problem
• The End of Handwritten Network Configs?
• What Agentic Search Means for Normal Humans
• Why Companies Are Already Abandoning LLMs