Which AI Medical Scribe Should a Hospital Pick Now That Epic Has Shipped Its Own?

Epic's tool already shipped. That changes the question.

In September 2025, the frame was future tense: Epic — EHR vendor to roughly 42% of US hospitals, and Abridge's own former shareholder — had announced it was building a Microsoft-powered ambient scribe, with a "limited release" pencilled in for early 2026 (Business Insider; STAT). A year later, at Epic's August 2026 Users Group Meeting, CEO Judy Faulkner said the wait was over: Chart with Art, Epic's native ambient-charting tool, is "in use by doctors, nurses and staff" and generating visit notes automatically, with Art-generated summaries now live at more than 300 healthcare organizations (Becker's Hospital Review). Epic has also extended it to inpatient nurses — Miami's Mount Sinai Medical Center became the fourth health system to roll it out to nursing staff in July 2026 (Becker's).

So the premise of "which scribe should a hospital pick before Epic ships one" is already out of date. Epic's own tool exists, it's in production, and it's running at roughly the same scale — 300-plus organizations — as Abridge's enterprise footprint. The real question for a hospital deciding now is narrower: given that Epic's native option is already there, does a health system still need to pay for a third-party scribe, and if so, which one.

What Chart with Art is, and what it isn't

Chart with Art listens to the visit and drafts the note inside Epic, with no separate login, integration project, or procurement cycle — it rides on the EHR contract the hospital already has. Epic hasn't published a separate price for it, which is itself informative: unlike the third-party vendors below, none of whom hide their pricing, Epic treats this as a feature of the platform rather than a product with its own price tag.

What it doesn't do, at least not yet, is the second half of what the leading independent scribes now sell: evidence-linked clinical decision support, coding-accuracy guarantees, and multi-EHR portability. Epic's parallel "Ergo Visit" framework adds pre-visit chart summaries and a Cosmos-powered feature called Best Care Choices for My Patient, which compares a patient's case against Epic's 320-million-patient deidentified database — but that's decision support built for Epic customers specifically, not something a hospital can take with it if it ever changes EHRs (TechTarget).

The two vendors actually built to survive Epic doing this

Abridge and Ambience Healthcare are the only two independent scribes that have kept pace with Epic's rollout at genuine hospital-system scale, and they've responded to Epic's move in opposite but coherent ways.

Abridge has doubled down on being the clinical-intelligence layer rather than just the note-writer. In August 2026 it opened its evidence-linked clinical decision support — pre-visit summaries, chart-grounded answers to clinical questions, referral drafting — to every clinician at its partner systems, not just the ones dictating notes through it. More than 300 enterprise health systems, covering over 250 million patients, have adopted the expanded tool since its April 2026 launch, and Abridge says monthly active users now cover more than half of eligible clinicians (Becker's). It's still the vendor with the deepest bench of named enterprise references — Kaiser Permanente's 25,000 clinicians across eight states (a system it won by displacing Nabla in 2024), plus Mayo Clinic, Johns Hopkins Medicine, Duke Health, and Yale New Haven Health (Business Insider). Notably, most of those are also Epic customers — Epic's own native tool hasn't dislodged Abridge from its biggest accounts, at least not yet.

Ambience Healthcare has taken the opposite bet: build deeper into Epic rather than around it. Days before Epic's own scribe announcement in August 2025, Ambience shipped an Epic-specific EHR-navigation agent built on OpenAI's GPT-5 (STAT), and its pitch to hospitals now centers on coding integrity rather than transcription alone — the company claims 95% coding compliance as verified by AAPC and a KLAS Trailblazer award for 2026, and cites Ardent Health returning "$3 of operating margin for every $1 spent" on the platform (vendor claims, ambiencehealthcare.com). It closed a $243 million Series C in July 2025, the largest recent raise in the category (Business Insider).

The practical distinction for a hospital: Abridge is the safer default if the priority is breadth of adoption and a track record with health systems your size — it has more named large-system references and now works across non-Epic EHRs too, including Cerner, Athenahealth, and Meditech. Ambience is the stronger pick if the CFO's chief concern is documentation feeding revenue-cycle accuracy rather than clinician time saved, since coding and audit compliance is where it has put its evidence.

Where Nabla lost the moment

Nabla is the cautionary case. It held Kaiser Permanente's ambient-scribe contract until Kaiser moved the entire 25,000-clinician system to Abridge in 2024 (Business Insider). In July 2026, Nabla installed a new CEO, Brian Manning, and its own leadership admitted the company hadn't built a US brand or go-to-market motion to match its rivals — "we really haven't accelerated our go-to-market in line with what others are doing," Manning told STAT, adding the company is now betting on "world models" as an LLM alternative to differentiate technically (STAT). Its last disclosed raise, a $70 million Series B in mid-2025, is an order of magnitude behind Ambience's Series C and Abridge's roughly $700 million raised to date. For a hospital picking now, Nabla is the vendor betting on a comeback, not the one with the deployment record to point to today.

The smaller vendors aren't built for this decision

Suki and DeepScribe both integrate with Epic and other major EHRs and serve real customer bases — DeepScribe cites 1,500+ healthcare organizations, mostly ambulatory and specialty practices rather than large inpatient systems. Sully.ai claims 30,000+ providers and bundles scribing with scheduling and front-desk tasks, a broader but shallower play than Abridge's or Ambience's enterprise clinical-intelligence pitch. Freed is priced and packaged for individual clinicians and small groups — $39 to $119 per clinician per month, custom pricing only at the "Groups" tier (getfreed.ai/pricing, verified today) — which makes it a reasonable choice for a solo practice or small clinic bolting AI onto Epic or another EHR, but not the enterprise procurement a hospital system is running.

Two other listings in this category, Atropos Health and Hippocratic AI, aren't scribes at all — Atropos generates real-world evidence answers to clinical questions, and Hippocratic AI builds voice agents for patient-facing conversations like intake and follow-up calls. Neither competes on documentation, so they don't belong in this comparison.

The actual answer

If your health system is already on Epic and hasn't signed a scribe contract yet, the first move is to actually turn on Chart with Art — it's very likely already available under your existing agreement, and for many specialties it may be enough. The gap you're deciding whether to pay to close is evidence-linked clinical decision support and audited coding accuracy, which Epic's native tool doesn't yet match.

If that gap matters — if the health system wants a vendor with a multi-year track record across the largest US systems and decision support that reaches every clinician, not just the ones dictating notes — Abridge remains the safer, better-referenced choice. If the priority is specifically revenue-cycle accuracy and coding compliance, and you're comfortable with a faster-growing but less-tenured vendor, Ambience is the stronger fit, particularly since it has chosen to build alongside Epic's ecosystem rather than compete head-on with it.

Nabla is not the pick right now unless your system is prepared to bet on its "world models" roadmap over its current deployment record. And nothing on this list should be evaluated before checking what Epic's own Art tool already does inside your build — the vendor comparison only matters for the capability gap that remains after that.