Three AI Shifts Changing How Health Care Decisions Get Made In Washington
Nearly three-quarters of U.S. employers plan to build AI into their health and benefits programs within two years, according to a 2026 WTW survey summarized by WorldatWork. That changes how employees find care, how plans administer benefits, and who reviews a denial. Here is what the Washington health care community should be asking about now.
1. AI is becoming the front door to benefits
Employers are putting AI in front of employees to answer benefits questions, steer people to care, personalize communications, and run administrative work.
Employers trust AI most for data analysis, communications, and administration. They pull back when it touches diagnosis, treatment recommendations, or chronic-condition management, according to the Business Group on Health’s employer survey.
The upside is real: complex benefits get easier to understand, and employees spend less time hunting for care. The test is whether the guidance is accurate, whether employees know they are talking to a machine, and how fast they can reach a person instead. Those three questions are worth asking any vendor directly.
2. AI is moving into clinical care, not just the back office
AI is already sitting alongside clinicians in imaging, cardiac monitoring, pathology, and screening. It flags what a reader might miss, sorts what to look at first, and shapes what lands in the chart. The FDA’s running list of AI-enabled medical devices cleared for U.S. marketing now spans radiology, cardiology, neurology, and other specialties.
As these tools enter routine care, a new set of questions follows. What does the plan cover. How does it pay. Which providers are equipped to use these tools well. Who answers when the tool gets it wrong.
The question worth sitting with is not whether AI replaces clinicians. It is that the same service, billed the same way, may now involve a different mix of human and machine judgment. Purchasers, plans, and providers all benefit from a shared way to see that shift and talk about it.
3. Washington wrote AI governance into law
In 2026 Washington enacted SB 5395. AI cannot be the sole means of denying, delaying, or modifying health care services. A licensed physician or other qualified health professional must review medical-necessity denials.
Plans using AI must also weigh an enrollee’s own clinical history and circumstances rather than leaning on group data alone, and must monitor for discrimination, accuracy, and privacy risks. RCW 48.43.830 sets the standards.
Starting October 1, 2026, certain carriers must report the share of prior-authorization denials that AI helped produce. That requirement sits in RCW 48.43.0161.
That reporting is where the conversation gets concrete. For the first time there is a number to look at instead of a reassurance.
Bring your questions on September 22
We are taking this conversation further at Future Health, Powered by AI — and Our Members, Tuesday, September 22, 3:30 to 5:30 p.m., Parkview Event Space in Seattle.
Employers, labor organizations, providers, and health plans will be in the room together, with a fireside conversation on what AI is actually doing in Washington health care and time to compare notes across sectors.
Your peers will be there. So should you. The event is open to WHA members, and if you are not sure whether your organization is one, reach out and we will sort it out before the 22nd.