Who Owns the AI Bill: A 2026 Buyer's Guide to Usage-Based Pricing for Clinical AI

Before your hospital signs for a clinical AI tool that bills by usage, decide who receives the invoice and what happens on the day it runs past budget. If nobody can answer both, the first month of adoption is also the first month of an overrun.

Becker's Hospital Review reported on October 16 that some AI tools charge by tokens, the small units of text a model reads and writes, instead of a flat license. Spending gets harder to predict as more people use them. At Children's Nebraska, Ryan Cameron, the chief information and innovation officer, told Becker's that rising usage costs could push clinicians to weigh expense against patient need. He argued that AI spending belongs to clinical, operational, IT, finance and governance leaders together, not to the CIO alone.

A flat license hides the cost of success. A meter shows it. That is useful information, and it is also a budget line that grows with every clinician who decides the tool is worth opening.

Three pricing models and how each one fails

Per-seat license. The price is fixed per named user. The risk is paying for people who never log in.
Usage-based (tokens, queries or documents). The price follows use. The risks are an invoice nobody can forecast and a tool that slows or stops when a cap trips.
Hybrid. A base fee covers an included volume, with overage above it. Predictable until the first heavy month.

The same Becker's story describes systems managing the meter in different ways. Jefferson Health matches tools to tiers of use. Advocate Health uses the cheapest reliable model for each task. Ochsner treats consumption pricing as a way to hold tools to a return on investment.

Six questions to ask before you sign

The six questions

What is the billing unit, and can a clinician explain it?Ask for the unit in plain words and a worked example priced from your own note or message volume. A vendor who cannot convert a token into one clinic session has handed you a number nobody in your building can forecast.
What does a heavy month cost?Ask for the bill at one, three and five times your expected volume, and for the range the vendor has seen at a customer of similar size.
What happens when a cap trips?The options are an alert, a hard stop, a silent slowdown and an overage charge. A tool that stops answering at 2 PM in the middle of a clinic is a different product from one that sends an email. The contract should name which one you get.
Who sees the usage report, and how often?Weekly, by department, to a named owner. A report that goes to a shared inbox goes to no one.
What can change without notice?Ask whether the vendor can change the model underneath you, change the price per unit, or both, and how many days of notice you get before either.
Who pays when the tool is working?Decide before adoption whether usage is charged back to the department or absorbed centrally. The units that adopt fastest are the ones whose budgets get surprised first.

Want the next buyer's guide in your inbox the morning it publishes? The Sarah Matt Briefing carries one every Tuesday. Subscribe here.

Which pricing fits which tool

A documentation tool that runs on every encounter has predictable, high volume. A seat license or a capped bundle usually fits.
A reference tool a clinician opens a few times a shift suits usage-based pricing, as long as the cap behavior is written down.
A pilot with unknown adoption fits a hybrid with a low base fee and a ceiling that stops the billing, not the service.

Where to start this quarter

List every AI tool in your hospital that bills by usage. Next to each, write the name of the person who sees the invoice and the date the budget resets. Where the second column is blank, the cap is whatever the vendor decided.

If the name in the first column is the CIO, ask the CIO one thing.

Who tells the unit when the meter runs out?

Subscribe to The Sarah Matt Briefing, one decision guide for health system and physician leaders every Tuesday. Subscribe here.

Clinical AI Governance Assessment

Dr. Sarah Matt is a surgery-trained physician-executive currently practicing internal medicine (charity care). She advises health systems and digital health companies on clinical AI implementation, vendor contracting and operational strategy.

Book a scoping call