The AI Portfolio Dilemma: Why Most Companies Can't Account for What They're Spending | CDO Matters Ep 104

Agents are about to hit data systems and governance frameworks at a volume, velocity, and variety they were never designed for. Bolting AI onto the old model may turn out to be the riskier bet. 📌 In this episode: Why agents bring back the three Vs of big data, and why even Google-scale systems weren't built for that much concurrency Winfried's first-principles view of governance: negotiation, direction, and accountability Why longer guardrail lists make agents worse at following rules, and why "human in the loop" is too passive a role Why data governance belongs inside corporate governance, not beside separate AI, IT, and risk frameworks 💬 The takeaway: "So much of what we've built over the last thirty years has become so monolithic and so big and so inflexible... I think we may be better off just starting from scratch." — Malcolm Hawker About the host + guests Malcolm Hawker is a former Gartner analyst, Chief Data Officer at Profisee, Editor-in-Chief of CDO Matters on Substack, and host of the CDO Matters Podcast. Guest: Joe Reis is [Owner, Data Engineer + Architect ] at ThinkQuick, LLC. Author of Mixed Model Arts, which rebuilds data modeling from first principles for both humans and AI agents. Guest: Winfried Etzel is [Sr Specialist Data Management and Podcast Host with DAMA Norway] at Equinor. Author of Data Governance in the Wild, which rethinks governance from first principles. Subscribe to CDO Matters Monthly → https://profisee.com/cdo-matters-community/#join-community Follow Malcolm on LinkedIn → https://www.linkedin.com/in/malhawker/ Guest → https://www.linkedin.com/in/josephreis/ Guest → https://www.linkedin.com/in/winfried-adalbert-etzel/

As companies hand agents more autonomy, the old idea of data ownership isn't enough, someone still has to be accountable when a machine gets it wrong. Kelle O'Neal, founder of First San Francisco Partners, joins Malcolm Hawker to talk through what "governance as code" actually looks like: RACI matrices for agents, formalizing data decision frameworks into AI decision rights, and the emerging insurance question nobody's asking yet. If your company has given AI a mandate but never named who's accountable, this one's overdue.

Every data team wants to be seen as a value driver instead of a cost center — but most still run like a service desk, taking orders instead of solving problems. Malcolm Hawker sits down with product consultant Nick Zervoudis to unpack the real difference between a product owner and a product manager, why "data democratization" isn't automatically good, and how a Qlik-funded study quietly invented the modern data literacy movement. If you've ever built a dashboard nobody opened twice, this conversation is for you.

Data leaders love to complain that the business doesn't get governance. At Davos this year, Global Data Strategy's Donna Burbank heard the opposite: CEOs and board members are actively asking for it, data teams just aren't in the room. Donna joins Malcolm Hawker to unpack why "data literacy" is the wrong frame, how a semantic-pedanticism spiral kills credibility in front of executives, and what listening, humility, and concrete examples do instead. If you've ever lost a boardroom by explaining data quality dimensions, this one's for you.

Every CDO is being told their data needs to be AI-ready. Very few know what that actually requires. Andreas Blumauer, SVP Growth at Graphwise and founder of the Semantic Web Company, makes the case that the missing piece isn't another pipeline or governance framework it's a knowledge graph. Malcolm Hawker sits down with Blumauer to work through what knowledge graphs actually do, why vector embeddings alone fall short for enterprise AI, and how CDOs can start small and build toward an infrastructure that lets AI reason across data silos instead of just searching through them. Originally released August 2025.