Many healthcare AI startups are changing the way they charge for their technology, moving away from flat subscription rates and toward pricing linked to the financial impact of their tools. Investors argue this aligns incentives and proves ROI, but health system buyers remain cautious, favoring predictable costs and measurable results.
Reinventing the SaaS model
Some healthcare AI startups are rethinking the traditional SaaS model tying their rates directly to the financial savings or added revenue that their tools create for customers, pointed out one health tech investor.
“Business models are shifting from SaaS pricing to transaction pricing. There’s lots of companies that charge per successfully completed action — especially on the agentic side. I think that’s what’s getting customers to sign up — not feeling like they’re paying a lot of money without the ROI being tied to it,” stated Vig Chandramouli, partner at Oak HC/FT.
This type of pricing structure is becoming particularly common among companies building AI tools for administrative workflows like patient intake, scheduling and revenue cycle management. Chandramouli said that many voice AI and intake automation companies now charge customers per successfully completed task — such as scheduling an appointment, completing an intake workflow or verifying a patient’s insurance benefits.
Some examples of companies that do this include Prosper AI, which charges per completed call or workflow, as well as voice and intake automation vendors like MedCalls.ai and Intelliclinic, whose fees scale with usage or successful task completion.
Even when contracts are structured to look like traditional SaaS agreements, the underlying economics often still revolve around transaction volume and success rates, Chandramouli noted.
Vendors may set minimum usage commitments, but the bulk of revenue is driven by how many tasks its platform completes — a model that is spreading to more workflows, like fax intake processing and payer calls, he added.
But the model has risks for startups. Chandramouli cited underpricing as one example.
“If they price, for example, on a success basis, and they’re doing 10 actions and one of them is successful, that’s where you get screwed in the unit economics. On the per successful action basis, you make money, but when you’re doing 10 of them, your margins are a lot worse, right? There’s things like that where you just have to be really careful on how it’s being contracted,” he explained.
Poorly defined contracts around ROI, variable payments and expected volumes can quickly destroy a startup’s unit economics, especially in larger, flashier contracts, Chandramouli said.
The buyer side of the equation
Another investor is skeptical about AI startups moving toward transaction-based pricing, saying the model isn’t attractive to health system customers.
Rachel Feinman — senior vice president of innovation, ventures and digital solutions at Tampa General Hospital and managing director of TGH Ventures — said health systems prefer more predictable pricing structures.
“We want to be able to model out what an ROI would be for a particular solution, and then have fixed pricing, because frankly, we don’t necessarily want to share the upside,” she remarked.
Feinman said her team has received plenty of pitches from startups with ROI-based pricing models, particularly in areas like revenue cycle management and coding automation, where vendors often propose taking a percentage of the additional revenue or savings their tools generate.
“We haven’t been receptive to that as of yet. I don’t know if the market will force us in that direction or not. But as of right now, we really try to find a baseline multiple on whatever cost savings or revenue generation there will be, and then set pricing based off of that,” she explained.
While investors may build projections around revenue-sharing or percentage-of-savings contracts, Feinman noted that many health system CFOs resist those arrangements.
She encouraged investors to talk directly with their potential customers to determine whether these business models are actually feasible in the market.
Outcomes-based contracts
Joey Dean, senior director of enterprise solutions at consulting firm Healthcare IT Leaders, said the real challenge isn’t whether a startup charges per transaction or uses a flat subscription model — it’s proving that the AI technology actually creates value.
In his view, outcome-based contracts — where vendors and health systems agree on shared goals — can better align incentives than simple transaction-based pricing.
“Both parties can really define those desired outcomes together and share an interest in achieving them,” Dean stated.
He said health systems care less about the number of tasks completed and more about quantifiable results.
“The health system really doesn’t care much about how many scheduling transactions an AI agent processes. They really care about the no-show rates. Did those drop? Did the time to the next available appointment improve? Did patients actually get seen? That’s the outcome — and I think that’s where the pricing model will start to follow,” Dean remarked.
Instead of paying per completed task, both parties set specific goals upfront — like reducing no-show rates or speeding up patient intake — and tie payment to achieving those outcomes.
He thinks this approach puts more responsibility on AI vendors to demonstrate real impact. Companies can no longer rely on flashy technology or marketing claims — with the model Dean favors, the value must be measurable and contractually defined.
Healthcare AI startups that fail to link their platforms to concrete outcomes risk losing their credibility with health systems and falling behind in a market increasingly focused on tangible results.
Photo: MirageC, Getty Images
