CDO Matters Podcast

The episode surfaces a few genuinely useful ideas - strategic decommissioning, on-premise as the new cloud, and the near-universal inability to track AI spend as a discrete category - but these are diluted by lengthy throat-clearing, the host's personal anecdotes, and concepts (the VAIN loop, the portfolio quadrant) that are named but never concretely explained.

“Strategic decommissioning. You should double down on shutting things off because this is eating your innovation budget”

“on premise is the new cloud for various types of workloads”

The PE/financial-lens framing of data investment and the shared-services-of-the-2000s analogy for AI CoEs are genuinely fresh, but the episode also recycles familiar takes on data products, ROI measurement, and build-vs-buy without pushing any of them to a truly contrarian conclusion.

“everyone has their AI center of excellence, enablement centers... it's basically the shared services of the early 2000s all over again”

“architecture is a financial variable rather than a pure technology discussion”

Ivo Sokolov is a legitimate practitioner - managing director of a boutique consultancy, co-author of a domain-specific book, and a board member of the Institute of Management Accountants - with real banking and regulated-industry client experience; he is not a generic thought-leader, but he is also not an operator who has personally run a major data function at scale.

“I sit on the U.S. Institute of Management Accountants board”

“Most of my clients are banks”

The episode offers a handful of concrete data points - the $200-to-$7K token-cost jump, the host's sub-5% estimate for companies that accurately track AI spend, and the BCBS239 regulation and Lehman Brothers reference - but the majority of claims are unanchored generalities about 'many companies,' 'various industries,' and unnamed clients.

“out of your $200 dollars subscription, if you recalculate the token cost for certain use cases, all of a sudden it's 7K”

“my guess would be under 5%”

The host frequently answers his own questions, delivers multi-sentence monologues where a follow-up question should be, and offers no meaningful pushback on any claim; the guest's vague assertions about the VAIN loop's mechanics, the quadrant's 70 factors, and unnamed client successes all pass unchallenged.

“Could not agree more, could not agree more”

“Love it, love it”