Beyond the Mandate: Why AI Adoption Stalls on Data, Cost, and Compliance

Operators across regulated and legacy-bound industries say the hard part of AI is not adoption enthusiasm but data infrastructure, consumption costs, fractured compliance, and the absence of real value measurement.

Senior operators across finance, HR, cybersecurity, entertainment, IT, and translation described a widening gap between pressure to adopt AI and the plumbing needed to do it safely. Entertainment and festival leaders said decades of data locked in 1980s and 1990s systems block anything beyond simple tasks without a full rebuild. An IT sourcing manager warned that consumption-based AI contracts let a single user burn tens of thousands of dollars overnight, with governance lagging far behind. A PEO manager described AI eroding the billable core of consulting, with a third of some offerings under pressure as clients self-serve. A financial services consultant flagged banks building parallel AI governance instead of integrating it into existing risk frameworks, adding burden. Regulated verticals fracture strategy: one firm running a Claude pilot must ban the same tools in its government vertical. And a translations leader called token-volume metrics nonsensical, since heavy usage can produce no value. The common thread is that measurement, governance, and infrastructure trail the mandate to deploy. Executives should treat data readiness and cost controls as prerequisites, not afterthoughts. They should reprice expertise services around judgment, integrate AI risk into existing frameworks, and replace activity metrics with outcome-based value measures. Operators repeatedly stressed keeping human discernment in the loop and writing clear guidance on how, not just whether, to use AI.