Why AI Transformation Needs a Strategic Management Office

As health plans have moved to adopt AI solutions, most have done so by accumulating use cases. It is not unusual for a payer organization to catalog 200 to 300 candidate applications and to regard that breadth as progress. The leadership team then reaches for the structure it knows best, the Project Management Office. From our experience, we strongly challenge whether that traditional structure suits the task of adopting AI.

The prioritization problem

The difficulty begins with prioritization. Payer executives generally recognize that only a few initiatives, perhaps three or four, will meaningfully move a supporting P&L. The organization does not lack for ideas so much as a mechanism for distinguishing those that matter from the many the plan might undertake. A PMO does not supply this mechanism, because it is built to track the work it is given rather than to decide what is worth giving it and how to execute.

The flexibility problem

The second hurdle is operational flexibility. The PMO works well for a familiar class of problem, where the objective is set in advance and success means delivering a fixed scope on time and on budget. AI transformation rarely presents itself in those terms. The work is exploratory, the scope shifts as understanding develops, and value is realized only when a change in operation or workflow actually takes hold.

A structure built to execute, not observe

Unlike the PMO, the SMO is forward-deployed: rather than managing from a distance, it embeds capability, including AI agents alongside experienced operators, within the work itself.

What this moment calls for is a structure built for execution rather than oversight: a Strategic Management Office, a lightweight strategy-and-execution function that concentrates attention on the few initiatives that matter, maintains an operating rhythm, and holds realized value against what was forecast.

Business-led, not technology-led

Even so, no office can accomplish this alone. AI transformation has long left the station of being technology-led; it is fully business-led, which means the board and multiple executives must stay engaged. Only those who own the business verticals hold the authority to change how it operates, and most AI initiatives depend on exactly such changes. Where this SMO model has been applied, agent-supported workflows have absorbed much of the traditional coordination that would otherwise occupy several PMO staff per workstream, so direction and execution sit with the same embedded team rather than being split between an office that tracks the work and the operators who actually do it.

The right instrument for a different task

The PMO remains the right instrument for a particular kind of work. The transformation now underway is simply a different kind. Structure alone will not guarantee the returns from AI that an organization hopes for, but an office built to concentrate effort, to execute rather than just observe, and to keep realized value in view gives the other conditions of success a far better chance of mattering.

3Pillars Solutions, How We AI. A perspective on the operating model for AI transformation in health plans, drawn from 3Pillars engagement experience. © 3Pillars Solutions.