# How Memorial Hermann measured $37,500 in AI scribe value per ER doctor

By MarketScale Newsroom · Published 2026-11-21 · Healthcare on MarketScale
Canonical: https://www.marketscale.com/industries/healthcare/how-memorial-hermann-measured-37500-in-ai-scribe-value-per-er-doctor

> Memorial Hermann ER doctors on Ambience saw 1.7 more patients a shift. The $37,500 figure is gross, and the way it was measured is the part buyers can reuse.

## Key points

- Memorial Hermann's $37,500 per emergency physician is gross: the gap in work RVUs per clinical hour between AI documentation users and non-users, annualized over 160 ten-hour shifts, before any software or implementation cost.
- Divided out, that gain comes to roughly $234 a shift or about $23 per clinical hour, a practical yardstick to hold against any all-in per-physician cost quote.
- When a vendor puts fees at risk against outcomes, the comparison group, billing source and acuity adjustment become contract terms, because they decide what counts as a gain worth paying for.

One point seven. That's how many more patients Memorial Hermann's emergency physicians who used Ambience Healthcare's AI documentation platform saw per 10-hour shift than colleagues who didn't. Run through billing data, the gap became the headline number in a KLAS Research ROI validation report that Ambience announced on Nov. 3: about $37,500 in additional annual productivity value per physician.

Ambience is careful about what that number is. Justin Beeker, the company's head of value attainment, told Healthcare Finance News reporter Nathan Eddy that the $37,500 is gross value and not a net return. It was calculated before software, implementation and other costs, all of which vary by contract. Memorial Hermann has already acted on the pilot and extended the platform beyond its emergency departments into ambulatory care.

Oddly enough, that candor makes the figure more useful. A health system CFO or revenue cycle leader weighing an ambient documentation contract can't use the $37,500 to settle whether the tool pays for itself. It does supply a measurement recipe built from ordinary billing data, along with a clear account of what that recipe can prove and what it can't.

## A recipe built from billing data

Beeker walked Healthcare Finance News through how the number was built. The analysis took the difference in work relative value units generated per clinical hour between Ambience users and non-users, then annualized it on 160 ten-hour shifts a year, an assumption he called conservative. The billing records came as provider-level monthly summaries from Memorial Hermann's third-party emergency department revenue cycle partner. The comparison kept department and time period the same for both groups.

KLAS based its report on in-depth interviews with Memorial Hermann stakeholders about the outcomes and how they were measured. The health system drew on clinician surveys, operational data and the independent billing figures. It also compared Ambience users with two groups: physicians with no documentation help and physicians working with human scribes. That second group gives the evaluation a reference point beyond doing nothing at all.

Block Field

The annualization is where a finance team can run its own numbers. Divide $37,500 by the 160 ten-hour shifts behind it and the gross gain comes to roughly $234 a shift, or about $23 per clinical hour, the yardstick any per-physician software and implementation bill has to clear.

The shift assumption works in both directions. A group whose emergency physicians work more than 160 shifts a year would get a larger annual figure from the same hourly gap. A group that relies heavily on part-time physicians would get a smaller one. Converting the result to an hourly rate before comparing it with a vendor quote avoids carrying Memorial Hermann's staffing pattern into someone else's budget.

## Case mix, coding and causation

According to Beeker, the billing comparison did not individually adjust for patient acuity, staffing or physician tenure. He said the results show how Ambience users performed relative to non-users and should not be read as a causal estimate. The report covered productivity and coding outcomes. It did not directly assess documentation accuracy or clinical quality.

That caveat bears most on the 1.3% rise in work RVUs per encounter. Ambience's announcement describes it as more accurate capture of physician work, and that reading is plausible: a note that records an encounter's complexity more completely can support more work RVUs. A difference in case mix between users and non-users could also move the number, though, and an unadjusted comparison can't tell the two apart.

Block Field

The design also has real guardrails. The revenue cycle partner held both groups to the same coding compliance standards, and clinicians still review and approve their own records. Because users and non-users were compared in the same departments over the same months, the seasonal and site differences a before-and-after study would pick up are removed. Then there's the expansion itself. After the 15-month pilot with about 100 emergency department physicians, Memorial Hermann expanded the platform to more than 560 clinicians across emergency medicine and ambulatory care. That decision suggests the health system was satisfied with how the pilot went.

## Hours saved and visits added

Providence offers a second reference point, measured in a different currency. At a Nov. 16 panel hosted by the Federal Reserve Board and the Federal Reserve Bank of San Francisco, President and CEO Erik Wexler said physicians at the 51-hospital system report that ambient documentation saves them two to four hours a day, as Alan Condon reported for Becker's Hospital Review.

The peer-reviewed numbers are smaller. A study published in JAMA Network Open in May, which tracked 1,547 Providence clinicians using ambient AI, found reductions in note-writing time and after-hours documentation, according to Becker's Hospital Review. The gains per clinician were modest, though the researchers said they could add up to meaningful relief across a large system. Memorial Hermann's survey feedback points the same way on after-hours work: physicians said they got more documentation done during the encounter itself.

Side by side, the two systems are counting different things. Providence's headline number is time, partly self-reported. Memorial Hermann's is dollars from billing data, driven by extra visits and higher work RVUs. That difference sets a trap for anyone building a business case. If the documentation time a physician saves is exactly what lets them see 1.7 more patients, that hour appears once as relief and once as revenue, and counting both would book it twice.

Which measure to lead with depends on what a system is buying the tool to fix. For a crowded emergency department, throughput is the natural metric. For a physician group where burnout and after-hours charting are the pressing problems, time saved may be the measure that matters.

## Contract terms that define the gain

One more detail from Beeker turns measurement into a procurement question. Ambience puts part of its fees at risk against outcomes measured on the customer's own data. Under that kind of deal, choices like the comparison group, the data source and whether to adjust for acuity become contract terms, because they decide what counts as a gain worth paying for.

Memorial Hermann's evaluation gives buyers a template for those terms, and the limits Beeker described show where to tighten them. Christopher Langan, MD, CEO of Memorial Hermann Health System Physician Organization, said the effort started with physician workload and a commitment to measuring whether the technology delivered. A system planning its own pilot could build that discipline around three questions:

1. What is the all-in cost per physician, including subscription, implementation and internal staff time, and how does it compare with roughly $234 a shift in gross value at Memorial Hermann's rate?
2. Which comparison group and billing source will at-risk fees be measured against, and will the analysis adjust for acuity, staffing and tenure?
3. How will documentation accuracy and coding compliance be audited, since a productivity comparison doesn't measure either?

So far, the pilot has measured only emergency medicine. Memorial Hermann now runs the platform in ambulatory care too, and the next figure worth watching is a billing-data comparison from those clinics, which would show whether the $37,500 holds up in a setting with its own visit patterns.

## Sources

- [Memorial Hermann reports productivity gains from AI ...](https://www.healthcarefinancenews.com/news/memorial-hermann-reports-productivity-gains-ai-documentation) (Healthcare Finance News)
- [Ambience Healthcare ROI Validations 2026 | KLAS Report](https://klasresearch.com/report/ambience-healthcare-roi-validations-2026-reducing-documentation-burden-to-improve-clinician-experience-and-performance/3948)
- [New KLAS Research Report Shows Value of Ambience at Memorial ...](https://www.ambiencehealthcare.com/blog/new-klas-research-report-shows-clinical-and-financial-impact-of-ambience-healthcare-at-memorial-hermann)
- ['We're seeing it now': Providence CEO on the AI productivity boom](https://www.beckershospitalreview.com/finance/were-seeing-it-now-providence-ceo-on-the-ai-productivity-boom/) (Becker's Hospital Review)

Tags: Memorial Hermann, Ambience Healthcare, KLAS Research, Ambient AI documentation, Emergency department productivity, Providence

---
Source: MarketScale, https://www.marketscale.com/industries/healthcare/how-memorial-hermann-measured-37500-in-ai-scribe-value-per-er-doctor. Published for AI indexing and citation; cite the canonical URL. Site guide for agents: https://www.marketscale.com/llms.txt
