Most health systems have deployed an ambient AI documentation solution by now, but many leaders haven’t asked the harder question: is the tool actually generating financial returns, or simply replacing dictation? Meanwhile, payers are accelerating their own use of AI in claims adjudication and denial management, widening the gap each quarter that documentation strategy stays in place.
During a featured session hosted by Ambience at Becker’s Hospital Review 16th Annual Meeting in April, three healthcare leaders unpacked what separates ambient AI deployments that move the financial needle from those that don’t. The panelists were:
- Felix Segre, director of revenue integrity at Memorial Hermann in Houston
- Michael Han, MD, clinical healthcare executive and former chief medical information officer at MultiCare Health System in Tacoma, Wash.
- Bala Hota, MD, independent informatics advisor and former vice president and chief analytics officer at Rush University System for Health in Chicago.
Below are four takeaways from their conversation.
Note: Quotes have been edited for length and clarity.
1. Surface adoption is masking the real value gap
The first generation of ambient scribes proved physicians would use them. The next phase is determining whether deployment is actually changing clinical and financial outcomes or simply sitting on top of existing workflows. Dr. Hota argued that surface-level adoption can mask whether providers are actually generating value, and that an intelligence layer — one that ties what’s said in the room to the correct ICD-10 code and the documentation that supports it — is what now distinguishes mature ambient AI tools from the rest.
2. Utilization has to be measured at the encounter level
Vendors talk about utilization, but the metric only matters when it reflects how often providers actually use the tool during a patient visit, not just whether they have a seat license.
“If you have 75% of your providers actually in a seat, but they’re only using it for a third of their encounters, you don’t have very high utilization,” Dr. Han said.
At MultiCare, which went live with an enterprise Ambience contract in January, about 65% of providers are using the tool for 83% of their encounters, according to Dr. Han.
3. Top-line revenue and cost to collect are a ‘both, and’
For Mr. Segre, the test of an ambient AI tool isn’t only whether it captures more revenue at the front end. It’s whether it reduces the rework that drives cost to collect at the back end.
“It has to be a song and dance,” Mr. Segre said. “Rework contributes into the cost to collect, and I’m very particular about that because if there is a rework and that solution is not addressing that, then that becomes a problem.”
He flagged three criteria any solution needs to meet: interoperability across the front, middle and back of the revenue cycle; accuracy that holds up against the underlying clinical data set; and usability that doesn’t create a heavy training burden.
4. The flywheel: clinician satisfaction feeds financial returns
Done well, ambient AI creates a self-reinforcing loop. Providers adopt the tool because documentation is accurate and reduces pajama time, accuracy at the source feeds clean coding, and clean coding reduces denials and back-office labor.
At MultiCare, Dr. Han reported a 6% productivity gain and an organic increase of about two patients per day among providers using the tool. Physician burnout dropped from 66% reporting at least one symptom to 16% — a 75% reduction — while the share of physicians reporting they enjoy their work climbed from 11% to 37%. Memorial Hermann’s early data showed a roughly 35% reduction in documentation time, from 121 minutes to 81 minutes, with encounter utilization of about 85%.
“It’s an unfair advantage for health systems that implement it,” Dr. Hota said. “If you do it the right way, you’ve built this whole cycle of better documentation, more accurate charts where it’s really the source of truth, and then you reduce back-office coder, CDI work, denials and having to file appeals.”
Bottom line for health system leaders
If revenue cycle and clinical leaders can’t say what percentage of encounters are being documented through their ambient AI tool, they can’t yet say whether the deployment is generating returns. With payers compounding their own AI advantage in adjudication, the cost of leaving “good enough” in place keeps growing.
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.
