Centene CEO: Health Care's Biggest AI Prize Isn't Billing Automation

Centene CEO Sarah London says most AI deployed at scale in health care today drives what she calls "inflationary use cases" — tools that generate more revenue in a fee-for-service system rather than reduce the underlying cost of care. She made the comments Monday in an interview at the Fortune Most Powerful Women Summit in Santa Barbara.

London runs a $33 billion company. Centene operates Medicaid plans in 29 states alongside marketplace and Medicare Advantage coverage, serving 26 million members. Those members, London said, are "among the poorest, sickest" people in their communities.

That population is often left behind by health care innovation. London sees two openings for AI to change that — and both sit well beyond the back office.

What's wrong with today's AI use cases?

AI can make billing faster, automate paperwork and help providers capture more revenue. But London argues those gains don't necessarily make health care less expensive. In a fee-for-service model, tools that help providers bill more efficiently can actually push spending higher, not lower.

The distinction matters for Centene specifically. Its 26 million members are Medicaid, marketplace and Medicare Advantage populations — groups where margins depend on managing the total cost of care, not maximizing reimbursements.

Where does London see the real AI opportunity?

She points to two concrete plays. The first is moving care into the community. In rural areas where medical services are scarce, AI could help community health workers triage non-acute problems locally — potentially avoiding emergency room visits altogether.

The second is patient control. Interactive tools could help people navigate initial decisions about a health concern without making every next step dependent on a clinician visit. That shifts routine decision-making away from expensive clinical settings.

"That's actually how you bend the cost curve," London said.

Does this make administrative automation worthless?

No — London says it sharpens the financial case for it. Every dollar saved on administration flows back into the product itself.

"Every dollar that we can save in the administration of health care, we can apply to benefits," London said.

For a company covering some of the country's most vulnerable patients, that trade is direct: administrative savings become richer benefits for people who can't absorb cost sharing.

What does this mean for health care CFOs?

The challenge, per London, is telling two very different AI investments apart. One tool speeds up billing. Another helps patients get the right care in the right setting — and that second category can change the economics of care itself, not just the speed of its paperwork.

The question isn't just how much AI can help health care do. It's whether AI can help health care spend less to deliver better care. For payers serving low-income populations, London's framing suggests the next procurement cycle will separate tools that inflate revenue from tools that genuinely cut utilization — and only one of those bends the cost curve.