
Episode #20
Build the Road Before you Build the Car
Two-thirds of senior executives can't point to a single measurable productivity gain from their AI investment. A quarter have already paused or abandoned a deployment outright. According to Ollie O'Donoghue, the reason isn't the technology but the road underneath it. In this episode of Stay Cognizant, host Pamela Axtell is joined by Ollie O'Donoghue, Head of Research at Cognizant, who leads the team behind Closing the Gap, a study of more than 1,100 senior leaders at G2000 companies across ten industries. Ollie spent years as a cloud analyst before taking over research, watching infrastructure go from a strategic conversation to an assumed commodity, and he makes the case that the assumption is now costing the G2000 roughly $2.4 trillion in revenue and cost savings it will never see. If you want a grounded read on why your pilots stalled and what the organizations ahead of you did differently, this episode gives you the data behind it. Together, Pamela and Ollie dig into what actually determines whether AI investment converts: the ten dimensions of technical infrastructure and the two that carry the most weight; why a single dimension slipping from good to adequate costs you 1 to 1.5% of your productivity outcome; and the four-quadrant view of where your organization really sits, including the finding that surprised the research team most, that focused investment outperforms broad investment even when your foundations are weak. Ollie also explains why compute is simultaneously the biggest inhibitor leaders name and the first line item they cut, what external partners bring that can't be built in-house, and what the organizations at the top are doing with the $1 to $2 billion the rest of the market is leaving on the table. Adequate used to be a passing grade. It isn't anymore and the enterprises seeing real returns are the ones that did the boring work first. Love the podcast? Don’t forget to subscribe and rate us! Timestamps: 0:00 Intro – Welcome to Stay Cognizant 2:00 $4.5 Trillion of Capability, Nothing to Show for It 5:00 Why AI Programs Start, Stall, and Get Scrapped 8:00 The Quadrant: Focused vs Broad Investment 11:00 A Formula 1 Car on a Dirt Road 12:00 Ten Infrastructure Dimensions and the End of Adequate 15:00 Compute Is the Weakest Link and the Hardest Business Case 19:00 Who's Ahead by Industry and Why Regulation Holds It Back 22:00 What an External Partner Brings That You Can't Build 24:00 Boring Is Cool and the Next Two Years Are About Results

