AI ROI Series | EP32 | Where's the AI ROI? | A Five-Step Guide to Measuring It
How do you measure AI ROI in 2026? Paul Brzozowski, Co-Founder of Olakai, the Enterprise AI System of Record, goes back to the roots of the AI ROI Series and reads four big reports side by side, from McKinsey, Wharton and GBK Collective, Deloitte, and Kyndryl. Most companies now say AI pays off. Kyndryl puts it at 64% of 3,700 leaders. Very few can say where. Gartner says only one in five AI use cases delivers a positive return, and McKinsey's group that can find AI in its profit has sat at about 6% for a year. In this episode, why hours saved no longer holds as a measure, what enterprise AI actually looks like across assistive, coding, and agentic AI, a real cost per finished task from OpenAI, and five steps to take control of your own AI ROI. The math is directional. Different studies ask different people different questions, so check it, and tell me where you read it differently. Chapters 0:00 Where's the AI ROI? 0:56 Four AI ROI reports, McKinsey to Kyndryl 2:18 Why your own numbers matter most 2:57 From 95% zero return to 64% saying yes 3:40 They say AI pays. Can they name it? 4:50 Why hours saved won't hold in 2026 6:02 How do you measure AI ROI per deployment? 6:58 The four kinds of enterprise AI 8:40 A beautiful, expensive mess, and AI FinOps 9:41 What does a finished task cost? OpenAI's numbers 11:11 Five steps to take control of AI ROI 12:39 An AI central bank, or a system of record 13:46 The line for the board New episodes every Tuesday (AI ROI Series) and Friday (Enterprise AI Weekly). Subscribe at https://www.youtube.com/@olakai-ai Read the full guide on LinkedIn https://www.linkedin.com/pulse/ai-roi-series-ep32-wheres-five-step-guide-measuring-paul-brzozowski-7lb7e/ Want to see your own cost per finished task? DM me on LinkedIn at https://www.linkedin.com/in/paulbrzozowski/ Learn more about Olakai, the Enterprise AI System of Record, at https://olakai.ai Sources. McKinsey, The State of AI in 2026, August 2026. Wharton Human-AI Research and GBK Collective, The Adaptive Advantage, October 2026. Deloitte Australia, 9 Nov 2026 (Australian results from a 3,500-leader global survey). Kyndryl AI Readiness Report, 10 Nov 2026. Gartner, IT Symposium/Xpo keynote takeaways, 2026. MIT NANDA, The GenAI Divide, August 2025. FinOps Foundation, State of FinOps 2026. OpenAI, 4 Nov 2026, company-reported. Per-unit costs are Olakai arithmetic and directional. Your AI is an investment. Let's measure it like one.


