The Hidden Cost of AI That Never Shows Up on the Invoice | EP46

AI token costs just changed — and the invoice is the least of your problems. Juan Orlandini, CTO of North America at Insight, breaks down what metered AI pricing is actually exposing about the structural assumptions your organization made when tokens were cheap or free. You'll walk away knowing how to read the real damage — the hiring decisions, workflow expectations, and delivery timelines that were built on subsidized pricing — and what to do about it before it compounds. Juan covers model routing, on-premises inferencing, token attribution, and why the shift to metered pricing is, counterintuitively, good news for organizations willing to act now. Read Juan's companion piece, "The Bill Was Never the Real Cost," for a deeper look at the structural risks hiding beneath your AI spend: https://www.insight.com/en_US/content-and-resources/blog/the-bill-was-never-the-real-cost.html Get Insight's AI and data expertise because you'll see exactly how to build a cost-disciplined AI strategy that compounds value across every project: https://www.insight.com/en_US/what-we-do/expertise/data-and-ai.html Subscribe to Insight On for more conversations with the technologists and leaders shaping enterprise AI. #AITokenomics #EnterpriseAI #AIStrategy Chapters 02:02 — The driveway analogy: when the bill hides the real damage 07:16 — How metered pricing exposed broken AI assumptions 08:22 — Model selection and the 4,500x price difference 09:21 — On-premises and on-device inferencing as a cost lever 11:09 — How to explain token economics to business leaders 12:03 — Token attribution: knowing if your AI spend is returning value 12:24 — What to do first when the AI bill shocks you 18:53 — Lessons learned for organizations building agentic systems today 24:28 — Red light, green light: five AI spending behaviors rated