Who Gets Paid for the AI Scribe? The wRVU Arbitrage Nobody Is Negotiating (2026)

Every physician publication has covered AI scribes from the burnout angle — and the burnout data is real. What no one has covered is the compensation angle: the arbitrage between what ambient AI documentation saves and who is entitled to the savings. This is not a hypothetical future problem. Researchers at Stanford, writing in a JAMA Network Open editorial in January 2026, explicitly cautioned health systems against exactly this move — urging them to emphasize "improvements in clinician turnover and well-being, rather than using efficiency gains to pivot to even higher productivity targets." When academic researchers feel compelled to warn health systems not to do something, it is because the incentive to do it is obvious and enormous.

This guide covers what the peer-reviewed evidence actually shows about AI scribe time savings (which is messier and more interesting than the vendor marketing), the precise math of what reclaimed documentation time is worth in wRVU terms, the three mechanisms — one loud, two silent — by which employers capture that value, the simultaneous 2026 CMS wRVU "efficiency adjustment" that is already extracting assumed productivity gains from physician pay, and the specific contract language physicians should be negotiating right now, before the first post-AI-deployment contract renewal cycle arrives and the norms get set without you.

What the Evidence Actually Shows About AI Scribe Time Savings

Before any compensation math, honesty about the data — because the gap between vendor marketing and peer-reviewed reality is itself part of this story, and because an article that inflates the numbers is an article nobody should cite.

The vendor claims: AI scribe companies and their marketing content routinely cite 2 to 3 hours of documentation time saved per physician per day, with practices seeing 15 to 25 percent more patients. Microsoft publicized "5 minutes saved per clinician per encounter on average" for DAX Copilot.

The peer-reviewed reality is more modest — and more nuanced:

• •The largest study to date — 1,800 clinicians across five academic medical centers, published in JAMA in 2026 — found AI scribe users saved approximately 16 minutes of documentation time and 13.4 fewer minutes in the EHR per eight hours of patient care, with 0.49 additional visits per week and roughly $167 per month in additional revenue per clinician.
• •A UCSF cohort study published in JAMA Network Open (January 2026) found AI scribe adopters generated 1.81 more wRVUs per week — a 5.8 percent increase — and handled 0.80 more patient encounters per week (2.8 percent), translating to approximately $3,044 in additional annual revenue per physician at 2025 Medicare rates, with no increase in claim denials.
• •A randomized controlled trial at UCLA across 238 physicians in 14 specialties found one tool (Nabla) reduced time-in-note by 9.5 percent while another (DAX) showed no statistically significant effect — with a critical contextual finding: the tools were used in only 30 to 34 percent of visits, and roughly 15 percent of assigned physicians never used their scribe at all.
• •A longitudinal study of 220 primary care clinicians across more than 314,000 encounters, reported by MGMA, found the most important pattern of all: documentation time savings grew from about 7 percent on day one to 15 percent by day 150. The benefit compounds with sustained use.
• •Individual health system deployments report larger numbers as adoption matures: Cooper University Healthcare measured 4.15 minutes saved per patient — approximately one hour or more daily; Intermountain Health saw a 27 percent reduction in time-in-notes per appointment among consistent users.

The honest synthesis: current average savings across all users in rigorous studies are modest — minutes per day, not hours — but savings among consistent, mature users are substantially larger and still growing as the technology and physician workflows improve. The burnout benefit, meanwhile, is unambiguous: a JAMA Network Open study of 263 physicians across six health systems found burnout dropped from 51.9 percent to 38.8 percent after just 30 days of AI scribe use.

Why this trajectory matters for the compensation argument: the question is not what AI scribes save today, averaged across half-hearted adopters. The question is what they will save in 2 to 3 years — the exact window of your next contract term — when adoption is consistent, the tools have improved, and automated coding, order entry, and inbox management (all actively shipping now — Ambience launched automated coding in February 2026) stack on top of pure note generation. You are not negotiating for today's 16 minutes. You are negotiating for the trajectory.

The Math: What Reclaimed Documentation Time Is Worth in wRVUs

Here is the calculation no AI scribe vendor puts in their ROI deck for physicians — because their ROI deck is written for the health system that pays the subscription, and the health system's ROI is this arbitrage.

The core equation: reclaimed documentation time can be converted into additional patient slots. Additional patient slots generate wRVUs. wRVUs have a dollar value — both to you (if you are above your productivity threshold) and to your employer (professional fee collections, plus downstream referrals, imaging, labs, and facility revenue that dwarf the professional fee).

Three scenarios, using a primary care physician on a $45/wRVU productivity model:

Specialty-side numbers run higher: a dermatologist adding patient volume at specialty visit rates, or a cardiologist whose added slots generate downstream echo, stress testing, and cath lab referrals, produces employer-side value multiples of the primary care figures. Industry ROI analyses cite scribe-enabled productivity gains of $99,000 to $220,000 annually per physician at full conversion, with one cardiovascular practice reporting over $1.3 million in added annual revenue.

Against these numbers, the AI scribe subscription costs the employer $99 to $600 per physician per month — $1,200 to $7,200 per year. Even at the conservative scenario, the employer's return on that subscription is severalfold. At the moderate scenario, it is one of the highest-ROI investments a health system can make.

And here is the arbitrage stated plainly: if the time savings are converted into added visits, and your compensation structure does not change, you capture the above-threshold wRVU value of those added visits — real money, and a fair split. But if the time savings are converted into added visits and your threshold rises to match, you capture nothing. You see more patients, generate more revenue, work the same total hours you worked before the tool arrived — and the entire value of the technology flows to the organization, while the tool continues to be described, in every internal communication, as a wellness initiative built for you.

The Three Capture Mechanisms: One Loud, Two Silent

Understanding exactly how the value gets captured matters, because only one of the three mechanisms is visible enough that physicians reliably notice it.

Mechanism 1: The Explicit Threshold Raise (Loud)

At contract renewal, the productivity threshold — the annual wRVU level above which your bonus activates — moves up. From 5,800 to 6,400. From the 50th percentile benchmark to the 60th. The justification offered, when one is offered, references "updated market benchmarks" or "operational efficiency initiatives." This is the visible version, and physicians at least see it happening and can push back with the benchmarking framework in our Physician Contract Negotiation guide.

Mechanism 2: The Silent Template Compression (Quiet)

No contract term changes at all. Instead, the scheduling template changes: 20-minute follow-up slots become 15; four half-day clinic sessions gain two slots each; the "standard" daily census expectation drifts from 18 to 22. Your threshold stays the same on paper — but the operational expectation of what a normal clinic day produces rises, and with it, the implicit baseline against which your productivity, your "citizenship," and your next raise are judged. Because nothing in the contract changed, there is nothing to formally object to — which is precisely what makes this the preferred mechanism. It requires no compensation committee approval and generates no negotiation event.

Mechanism 3: Benchmark Drift (Invisible — and the One Nobody Is Talking About)

This is the most important mechanism and the least understood, so it deserves full explanation.

Most employed physician contracts set productivity thresholds by reference to external survey benchmarks — MGMA median wRVUs for the specialty, updated annually. Roughly 70 percent of physicians now have a wRVU productivity component in their compensation, and threshold-set-at-the-median is the standard structure.

Now follow the logic forward: industry projections indicate AI scribe access will exceed 50 percent of providers by the end of 2026, on a path to becoming standard infrastructure within 3 to 5 years. As adoption spreads and matures, median physician wRVU production itself rises — because the survey population is increasingly AI-assisted. The MGMA median is not a fixed standard. It is a moving average of what physicians actually produce. A contract that says "threshold = MGMA 50th percentile, updated annually" contains an automatic escalator: as AI-assisted productivity becomes the norm, the benchmark rises, your threshold rises with it, and no one at your employer ever had to decide to capture your efficiency gains — the contract structure captured them automatically, industry-wide, with no fingerprints.

This is the mechanism that makes "my contract didn't change" a false comfort. If your threshold floats on an external benchmark, your contract changes every year whether anyone amends it or not.

The Double Squeeze: CMS Is Already Capturing Assumed Efficiency Gains

Here is where this stops being a future risk and becomes a present, structural fact — because a version of this exact capture has already happened at the federal level, effective January 1, 2026.

CMS finalized a -2.5 percent "efficiency adjustment" to work RVU values for nearly all non-time-based CPT codes — procedures, imaging, diagnostics, and most surgical services — affecting roughly 7,000 physician services representing over 90 percent of physician-billed Medicare volume. The premise of the adjustment is precisely the premise of this article: that physician work is becoming more efficient over time (technology being a central assumed driver), and that the wRVU values assigned to each service should therefore be marked down to reflect it. As compensation consultants at PYA have documented, the practical effect is that physicians performing identical clinical work in 2026 will register as less productive than in 2025 — potentially missing bonus thresholds and triggering guarantee reconciliations through pure measurement deflation, with no change in actual effort.

Put the two squeezes together and the picture sharpens: CMS is reducing the measured value of each unit of your work on the theory that technology is making you more efficient — while your employer, deploying the actual technology, is positioned to raise the number of units expected of you for the same pay. The physician sits at the intersection of two simultaneous extractions of the same efficiency gain: one from the payer side (devalued wRVUs), one from the employer side (raised expectations). A physician whose contract has neither a wRVU-value-neutrality clause nor a threshold-stability clause is exposed to both at once.

This is also why the Stanford editorial warning in JAMA Network Open matters so much as a citation: the academic literature has now formally identified "using efficiency gains to pivot to even higher productivity targets" as the failure mode that would squander the technology's burnout benefit. The researchers' explicit recommendation — that health systems should bank the ROI in retention and well-being rather than throughput extraction — is the strongest third-party support available for the negotiating position this article lays out.

The Contract Language That Doesn't Exist Yet — and Should

This is the practical core of this article, and to our knowledge, the first attempt anywhere to specify what AI-era productivity protections in a physician contract should actually contain. None of this is standard yet. That is exactly the point: the physicians negotiating contracts in 2026 and 2027 are setting the norms, and norms set in the absence of physician input will not favor physicians. Bring these concepts — reviewed by a qualified contract attorney or professional review service (see our Contract Diagnostics vs. Physicians Thrive comparison for options) — to any new contract or renewal.

The Questions to Ask the Day Your Employer Announces an AI Scribe

When the rollout email arrives — and if it hasn't yet, it will — these are the questions to raise, in writing, with your medical director or compensation committee:

1. Will productivity thresholds, panel size expectations, or scheduling templates change in connection with this deployment — this year or at renewal? Get the answer in writing. A verbal "no plans to" is not a commitment; see our Physician Contract Red Flags guide on why verbal assurances are worth exactly nothing at renewal time.
2. How will the organization measure this tool's ROI — and is added visit volume part of the business case presented to leadership? If the internal business case is built on throughput (it almost always is), the pivot to higher targets is already planned, whatever the wellness messaging says.
3. Will AI-assisted productivity data be used in individual performance evaluation or compensation decisions?
4. If my measured wRVUs rise after adoption, does that reset the baseline against which my future production is judged?
5. Who reviews and bears responsibility for AI-drafted documentation errors, and how is the review time accounted for? The UCLA trial found clinically significant inaccuracies occur "occasionally" — oversight is real work, and it is currently uncompensated work.

If You Own Your Practice, You Are the Arbitrage

One group of physicians should read this entire analysis in reverse: practice owners. For an independent physician or physician-owned group, there is no capture problem — the AI scribe's economics flow entirely to you. A $250 to $400 per month subscription against even the conservative-scenario value of one added visit per day is among the highest-ROI operational investments available in independent practice, with break-even typically reached within 2 to 4 months. Or — and this is the option no health system will ever offer its employed physicians — you can convert the reclaimed time into nothing at all: the same patient volume, shorter days, and a practice you can sustain for ten more years. The freedom to choose between those two uses of your own reclaimed time is, in miniature, the entire ownership argument made throughout this site's practice-model coverage, from our DPC and Concierge Economics guide to our Medical Practice Partnership Buy-In guide. Employed physicians negotiate for a share of the arbitrage. Owners simply keep it.

Frequently Asked Questions