Table of Contents
When an AI vendor walks into Mass General Brigham with a bold ROI claim, the conversation doesn’t start with a pilot proposal. It starts with a form, an early sign of just how structured AI vendor accountability healthcare practices have become.
Why Mass General Brigham Requires a Standard Assessment First
Rebecca Mishuris, MD, chief health information officer at Somerville, Massachusetts-based Mass General Brigham, requires every AI company to complete a standard assessment questionnaire before any procurement conversation advances. The form ties directly to the system’s responsible use of AI principles and covers how a model works, how it was validated, and whether the supplier looked for bias and equity issues during development. “Even being able to fill out that form is a non-negotiable,” Mishuris said.
A Broader Industry Trend Toward Structured Evaluation
That kind of structured scrutiny is becoming a defining feature of AI vendor accountability healthcare at health systems nationwide. As an increasingly crowded market produces increasingly confident claims, IT and informatics leaders say they are building more formal evaluation processes to hold suppliers accountable for what they promise and to share the risk when no one has proven results yet.
When Health Systems Must Build Their Own Evidence Base
When Mass General Brigham breaks new ground, it must sometimes build its own evidence base. The system ran what was at the time the largest ambient documentation pilot in the country, with 800 providers, and published the initial performance data where none had previously existed at that scale. In those circumstances, Mishuris looks for partners willing to share the accountability and risk, believing the arrangement produces better outcomes for both parties.
Why Shared Risk Leads to Better Products
“It leads to a better product,” Mishuris said. “It leads to us really understanding how to incorporate it into our environment, and it leads to that kind of precedent that the company needs to go talk to other organizations, and that we need internally to demonstrate that there’s value in scaling.”
Negotiating Failure Before Signing Any Contract
Kevin Bidtah, chief information and inclusion officer at Evergreen Health, a federally qualified health center with seven sites in Buffalo, New York, starts the accountability conversation before a contract is signed. Scope of work, expected outcomes, timelines, and the consequences of missed commitments are negotiated in writing. Then he asks a question these companies don’t always hear.
A Question That Sets the Tone for AI Vendor Accountability Healthcare
“When this fails, where are you headed? What are we going to do? How are we going to handle this failure?” Bidtah said. “Because not everything is going to just work right out of the box. Most things don’t.” His vetting process extends beyond the negotiating table, reaching out to peers who have worked with the same company, not to learn about features, but to understand what the working relationship actually looked like.
Why Go-Live Isn’t the Finish Line
That accountability standard holds for the lifecycle of the product’s use. “Go-live isn’t the end, isn’t the finish line,” Bidtah said. “It’s beyond that.” This ongoing accountability model reflects a broader shift in how health systems are approaching AI vendor accountability healthcare as a continuous relationship rather than a one-time transaction.
Drawing a Clear Line Between Companies Who Earn Contracts and Those Who Don’t
Stephanie Hines, CIO of Valleywise Health in Phoenix, draws a clear line between suppliers who can earn a contract and those who can’t. For any product that isn’t being co-developed with the health system, she expects these companies to show their work by outlining their training methodology, model maturation, and outcomes data from comparable organizations.
What Transparency Looks Like in Practice
“I want to see the evidence,” Hines said. “I want to see what the results have been. I want to see how you’ve trained your model. I want to see behind the curtain essentially of how you have matured your product.” A company that can’t provide that transparency is sending a clear message. “Anyone who has a bold claim around an ROI but they can’t articulate that information, they can’t share what’s behind that curtain, then that’s obviously a red flag,” Hines added.
What the Maturing Market Means for Evaluation
As the market matures, Hines expects performance to become easier to benchmark and weaker products to fall away. In the meantime, she sees shared accountability for outcomes, wins and losses, as the clearest signal a partner is worth a long-term relationship, a standard increasingly central to AI vendor accountability healthcare decisions industry-wide.
Why Stakeholder Buy-In Matters Beyond IT
Joseph Izzo, MD, chief medical information officer at San Joaquin General Hospital in French Camp, California, said the strongest presentations he has seen share one trait: the company doesn’t make the case to IT alone. They understand all stakeholders within the hospital’s ecosystem and can draw a clear value line for each.
Building Consensus Across the Organization
“Being able to actually sit down and have a company meet with our stakeholders, make those convincing arguments alongside me, I think it goes a long way,” Izzo said. This emphasis on cross-functional engagement suggests that effective AI vendor accountability healthcare practices extend well beyond IT departments, requiring these companies to build credibility with clinical, operational, and administrative stakeholders alike.
What This Means for the Future of AI Procurement
Across these four health systems, from academic medical centers to federally qualified health centers to safety-net hospitals, a consistent theme emerges: AI vendor accountability healthcare now depends on evidence, shared risk, and sustained transparency rather than confident sales pitches alone. As more health systems formalize these evaluation processes, companies unable to substantiate their claims with real validation data and comparable outcomes may increasingly struggle to gain traction, while those willing to share risk and demonstrate transparency stand to build the kind of long-term partnerships these CIOs describe as most valuable.
For more healthcare industry updates, insights and news, visit DistilINFO. Click here to subscribe to stay informed.
