AI Isn't Transforming Finance — And Healthcare Is Next in Line to Learn That Lesson

Healthcare is burning. So is finance. Both sectors have poured capital into AI tooling, and both are staring at the same uncomfortable result: productivity hasn't moved the way the models said it would. The instinct at most executive tables is to blame adoption rates or change management. The more honest diagnosis is that the organizations deployed AI into broken workflows and expected the technology to do the redesign work for them.

MIT Sloan researchers found that AI is failing to transform finance functions not because the tools are immature, but because leadership hasn't evolved the actual structure of financial work to accommodate them. The study, drawn from multiyear action design research across organizations undergoing digital transformation, identified a consistent pattern: AI gets bolted onto existing processes rather than used as a prompt to fundamentally rethink what those processes are for. The technology ends up accelerating the wrong things faster.

Healthcare is facing a structurally identical problem, at greater human cost. McKinsey's analysis of the future of work in healthcare argues that the sector is in a productivity crisis that more hiring and more technology alone will not solve. The next era of care depends on making automation work through human–AI workflows — not workflows where AI is an add-on to what clinicians and administrators already do, but genuinely redesigned operating models where the division of labor between human judgment and machine execution has been deliberately thought through.

That distinction — add-on versus redesign — is where most organizations are currently failing. And it's a leadership failure, not a technology failure.

Here's what makes this pattern dangerous: the two sectors are not outliers. They are bellwethers. Finance and healthcare are among the most data-rich, process-heavy, and compliance-constrained industries in existence. If AI can't produce structural productivity gains there, the implied premise — that knowledge work broadly will be transformed by current-generation AI — deserves considerably more scrutiny than most board decks apply to it.

The deeper problem is that executives are confusing deployment with transformation. Deploying an AI tool to summarize clinical notes or flag anomalies in a financial close is a workflow optimization. It may reduce time on task by 20%. It does not change the fundamental economics of care delivery or the cost structure of a finance function. Transformation requires asking which tasks should not exist at all, which decisions should be made at different levels of the organization, and which human roles need to expand in judgment and accountability as machines absorb the procedural load. Most organizations haven't started that conversation.

The industrial relations dimension compounds the risk. Personnel Today has raised the question of whether AI is becoming HR's defining industrial relations battleground — and the answer, increasingly, is yes. Unions and works councils are watching AI deployment closely. In sectors where collective bargaining is active, the absence of a coherent human–AI workflow story isn't just a strategy gap; it becomes a negotiating liability. HR leaders who haven't yet built a defensible position on how AI changes role design, not just headcount, are walking into those conversations unarmed.

What's missing from most transformation programs is a theory of the human role after automation. The McKinsey healthcare analysis is explicit about this: the productivity opportunity isn't in replacing clinicians, it's in removing the administrative burden that consumes an estimated significant share of their working hours, so that human judgment can be concentrated where it actually creates value. That's a workflow architecture question, not a technology procurement question. It requires operations executives and HR leaders to co-own the redesign — not hand it to the IT function once the vendor contract is signed.

The finance research from MIT Sloan points to a related gap at the leadership level: how leadership work itself evolves under conditions of technological and market uncertainty. Organizations that are getting AI deployments right are the ones where senior leaders have changed their own decision-making patterns, not just the tools their teams use. That's a harder ask than another digital transformation roadmap.

So here is the board-level question worth forcing: does your organization have a named executive accountable for workflow redesign — distinct from AI deployment — and have you tied their success metrics to productivity outcomes rather than adoption rates? If the answer is no, you are not behind on technology. You are behind on organizational design, and that gap compounds every quarter you don't close it.

The sectors that figure this out first won't be the ones that spent most on AI. They'll be the ones that were honest earliest about what AI cannot do on its own.