Every AI initiative has a budget and a builder. Far fewer have someone who answers for what it’s worth once it’s running. That gap sits at the center of the work Optura does with healthcare organizations, and it’s why Optura put the question of ownership in front of a cross-industry room of industry leaders at In the Round 2026.
Chief AI officers are spreading fast, with 76% of organizations now having one, up from 26% a year earlier. Yet when the Who Owns the Number panel asked the room to raise a hand if one person owned the value of AI, the answer was far from universal, and the panel itself didn’t agree on who it should be.
The case against the obvious owner
The obvious owner is whoever runs the process AI is changing. The argument against it is that AI combines processes and reimagines them, so today’s process owner may be the wrong person to own tomorrow’s version. Start with the outcome the organization wants, then find the person who can deliver it.
The case for the CEO
“We don’t have chief electricity officers and chief internet officers.” AI runs through everything, and that’s the case for the CEO owning it.
In the short term, it helps to have one curious person whose job is making people uncomfortable with being comfortable. At one technology company, every monthly all-staff meeting opens with someone from operations sharing what they built with AI. One of the company’s top three AI users came from quality management, and a middle-of-the-pack attorney built a system that triages incoming contracts and checks redlines against standard responses. The people who move AI forward show curiosity, agency, and urgency, and their titles don’t predict it.
The case for three in a box
One pharmacy benefit company tried three versions. The CIO and COO came first, and that produced a long list of small, scattered efforts. Then the business took the lead, aligned with the board. Now a business leader, a technology partner, and HR own it together, because most of the work turned out to be change management rather than technology. The company has also committed to its board to deliver 20% more productivity over three years, owned by each business leader.
The case for an executive and a doer
Ownership tends to work in pairs. An executive owns the value, and a doer owns the execution. Whoever owns the number also has to understand the model well enough to stand behind it. For a clinical leader, that can mean trusting it with their license on the line.
What happens without an owner
A prior authorization tool at the same pharmacy benefit company saved 10% of the time on 60 pages of clinical documentation and made a great case study for leadership. The clinicians never fully bought in, and the organization couldn’t say whether it now needed 10% fewer people or could handle 10% more requests. The savings were real, and they had no owner.
Where the debate landed
For all the disagreement about who, the panel lined up on how.
Delegate, don’t abdicate. “When you delegate, you delegate a task. You don’t abdicate responsibility.” Agents need the same management as people, since they don’t behave the same way every time. One team learned that after an AI agent fired one of the agents it was managing, and responded with a control plane of expectations and checklists.
Review it in conversation. At one health plan, the executive AI review grew from a two-hour agenda slot to a monthly half day to every two weeks, with executives in the room. Teams stopped bringing decks, because rehearsed presentations had turned the review into a show. An AI review board on its own hadn’t worked.
Make every headcount request account for AI. One CEO approved no new headcount for any department until its leader presented an AI plan. A team that asked to grow 30% ended up needing only to backfill attrition.
Watch the spend. Teams at one organization walked through their AI usage costs, initiative by initiative, at 6 p.m. every day for a month. Over the next two years, the cost of their first initiative, drafting replies to patient messages, fell about 99%.
Count the right things. Start with the outcomes the organization wants, work back to the measures that show progress, and map those to spend. A balanced scorecard links AI to retention, sales, and new business, not only the short-term bottom line. For now, learning counts as value too.
The question of who owns the number doesn’t have one answer, but it does have a test. If no one can say who decided what AI was supposed to deliver, and who’s accountable for what happened, no one owns it. Optura helps health plans and health systems answer both questions from the business case to the P&L, so ownership is clear before the work starts and stays clear while it runs.
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.
