Seven AI Questions to Expect From Your Board

The boards now overseeing AI in healthcare have already watched hundreds of millions of dollars go into data infrastructure builds that delivered very little. Executives bringing AI to the board should expect sharper questions this time around.

Boards are where the value of AI gets decided, funded, or cut, which is why Optura brought this conversation to In the Round 2026. The Winning the Boardroom panel, a mix of healthcare and technology leaders, including board members, agreed on who does what. Operators drive AI decisions. Boards set the focus, the risk tolerance, and the long-term strategy, and decide how they want results reported. That puts the work on management to bring the board what it needs to govern. Here are the seven questions to prepare for, and what to bring to each.

Leadership teams are prone to groupthink. CEOs and their teams attend the same conferences and hear the same panels, and that sameness shows up in the boardroom as hand-waving and investments made because everyone else is making them.

Bring: the reason this initiative matters to your organization in particular, and where it fits in the strategy.

The relationship works best as a back-and-forth. Operators drive the work and bring what deployed AI has taught them back to the board, and those lessons shape capital allocation and strategy. The lessons aren’t always flattering. One organization deployed clinical documentation tools years ago and still isn’t seeing a productivity gain, though the hope now is that it will show up in staff retention. A board that hears that early can redirect the next dollar before more is committed.

Bring: what has worked, what hasn’t, and what it changed about where you’d put the next dollar.

The data builds that failed had no interim measures, so boards only saw the final result. Value also depends on whether the work around AI changes. With ambient documentation, for example, if scheduling doesn’t change to use the time it frees up, the value never shows up.

Bring: the expected time to value and the interim measures that will show early whether it’s working, so the board can pull back before the money is spent.

Boards want to know how disagreements with AI get resolved, and who’s liable when a person overrides its recommendation. Those answers belong in the design, before anything goes live. When one home-based care organization launched a patient-facing AI agent, it had to work through legal and compliance questions about how the agent identifies itself as AI. Many patients still didn’t realize they’d spoken with one.

Bring: who has final say, and how an override gets recorded.

AI can drift after it goes live, and boards want to know someone is watching. It’s right for a board to focus on what could go wrong, so expect this one early. Part of the answer is data. Hallucinations drop sharply when AI has the right data available.

Bring: who watches it, which measures trigger a review, and the conditions under which you’d stop.

Workforce readiness is one of the three areas boards focus on, alongside risk and compliance, and capital allocation and strategy. Some employees will refuse to use AI at all. Junior staff may feel they’re already behind and under pressure to be far more productive because of AI. The home-based care organization behind that patient-facing agent shows what a plan can look like. It was the organization’s first patient-facing AI agent to make it through internal governance. During testing, clinicians tried to make the agent hallucinate, and did, before it went live. Its engagement rate runs about a third of the human team’s, but it can make a thousand calls without new hires. The organization sunset its training team because AI now coaches staff, and those people moved into roles that need more human contact. Staff are now asking for more AI help, because it takes the work they least want to do.

Bring: the adoption plan, and how roles will change rather than disappear.

There’s no AI strategy without a data strategy. AI is only as good as the data it runs on, and revenue cycle AI depends on what’s in the EHR. Boards that get this right don’t ask which point solution to buy. They ask how AI purchasing fits together across the organization. One hospital system leader estimated that about 85% of the point solutions pitched to their organization are far beyond what it’s ready to consume. A learning agenda with a deadline, set before the organization commits to build, buy, or partner, keeps those choices deliberate.

Bring: the state of the data, how this fits the rest of your AI portfolio, and what you’ve decided not to buy.

A real example, not a principles deck. A high-level presentation of principles “puts all of us to sleep,” as one executive put it. Take one concrete example and work it until the team hits a block, then figure out how to solve the block. Starting with every possible risk only brings frustration and disillusionment.

A regular back-and-forth, not a one-time pitch. Educate the board along the way, with operators bringing what deployed AI has taught them.

Meeting time for strategy. AI-prepared materials, or board members using AI themselves, mean the board arrives up to speed and the meeting goes to strategy and ideas.

The hard conversations, early. Leadership changes, the infrastructure spending that has to come first, and what gets cut to fund AI belong in the discussion before the decision, not in the middle of it. The executive sponsoring the work also has to stay close to it. When the sponsor steps away, the work starts to stall.

“Healthcare doesn’t change until it’s forced to change,” one panelist said. Healthcare still runs on fax machines, while banks solved data exchange years ago. That gap is also the opening, since healthcare has the most to gain from AI and the clearest chance to leapfrog. The money is tightening at the same time. For the first time in decades, federal health policy isn’t adding money to the system, and employers are pushing back on costs. With thin margins, every AI investment takes from something else.

“Slapping AI on top of a dysfunctional system is just going to lead to more dysfunction and financial loss,” one panelist said.

In that climate, a board needs to see what each AI initiative was approved to deliver and what it has delivered so far, not a story about potential. Optura gives healthcare leaders a way to show that, with numbers a board can act on.

For the themes that ran across every panel and roundtable, read Full Measure: The Set List, Optura’s insights report from In the Round 2026.