The Plumbing Problem: Why AI Adoption Stalls Behind Legacy Data and Loose Governance

Operators across regulated and legacy-heavy industries are hitting structural walls in AI adoption: old data systems, consumption cost risk, fragmented governance, and metrics that measure activity instead of value.

Senior operators across finance, IT, HR, entertainment, credit unions, and professional services described a consistent gap between pressure to adopt AI and the plumbing needed to do it safely. Legacy data trapped in decades-old systems blocks meaningful AI use in several industries, leaving firms with only surface wins like email drafting. Consumption-based pricing has shifted cost risk onto buyers, who lack real-time monitoring while vendors ship AI-enabled tools ahead of governance. Regulated verticals ban the very tools other business units pilot, fracturing any single rollout, and banks compound the problem by building separate AI governance rather than folding it into existing risk frameworks. In regulated settings, manual data redaction can cancel out the efficiency AI promises. Professional services face direct revenue erosion as clients now do with AI the work they once paid consultants for. Boards push AI for competitive advantage, but chosen use cases often copy competitors instead of fitting the institution, and success is measured by token counts that reward volume over value. The through-line: enthusiasm is running ahead of data readiness, cost controls, governance, and measurement. Executives should treat data migration, consumption caps, vertical-specific policy, and value-based metrics as preconditions, not afterthoughts. Those who reposition expertise around judgment and oversight, and who integrate governance into what already exists, will move faster and spend less than rivals chasing headlines.