Our read
Start from your EHR, your size and your data rules, not the brand. On Epic with many specialties, test Ambience first: John Muir chose it because it covered every specialty with an Epic integration. Test Abridge if you want rollout precedent, since Sutter and CHRISTUS both expanded after a pilot. In a community health center on athenahealth or MEDITECH, start with Suki. If note retention and hosting location decide, look at Nabla: its trust page says notes are kept 14 days by default and non-US clients are hosted in the EU. In Europe, Tandem is what your peers use, from Aleris to Rothschild. Corti is a developer platform, not a ready scribe.
Which tool for which job?
| Your job | Where to start | The reference to read | What it documents |
|---|---|---|---|
| Scale a pilot to the whole health system |  Abridge |  CHRISTUS Health | CHRISTUS Health ran a 2-month Abridge pilot, then signed an enterprise deal; its ambulatory CMIO says implementation took 8 days. 5 |
| Add patient-facing summaries to the record | Abridge |  Sutter Health | Sutter Health expanded Abridge to 1,000+ clinicians within three months of a pilot and built patient-facing summaries into its electronic health record. 2 |
| Document inside Epic with coding support |  Ambience Healthcare |  Onvida Health | Onvida Health runs Ambience inside Epic Hyperspace and Haiku for chart summaries, ambient notes, coding and patient instructions across 200+ specialties. 6 |
| Start with many specialties in Epic | Ambience Healthcare |  John Muir Health | John Muir Health deployed Ambience across 15 specialties after choosing it as the only option covering all its specialties with an end-to-end Epic integration. 1 |
| Roll out across specialties, with dictation in Epic |  Suki |  RUSH University System for Health | Rush expanded Suki enterprise-wide across 28 specialties after a trial and is co-developing dictation inside Epic with Suki. 7 |
| Document in a community health center on athenahealth | Suki |  Utah Navajo Health System | Utah Navajo Health System chose Suki; its CEO says integration with the athenahealth EHR was the clinicians' main differentiator. 3 |
| Roll out across dozens of clinics in Europe |  Tandem Health |  Aleris | Aleris rolled Tandem out across 60+ clinics in Sweden, onboarding 750+ clinicians who create 25,000+ notes a month, with weekly usage reports. 8 |
What does a hospital listed as a customer actually prove?
Read each hospital name by the page that carries it. A case study or press release (CONFIRMED) describes a deployment. A logo on the vendor's customer page (OBSERVED) shows only that the vendor lists the organisation. As of 2026-10-23, usedby lists 89 hospitals and care providers that use clinical documentation AI tools: 41 CONFIRMED and 48 OBSERVED. These are counts of named organisations, not market share and not a measure of quality.
Only some sources describe the work itself: pilot size, EHR, specialties. Those are the cases used below. Where a source names a hospital without any task, it supports adoption and nothing more.
Which scribe for an Epic health system: Ambience or Abridge?
If you run Epic and your clinicians span many specialties, start with Ambience. John Muir Health says on Ambience's site that it began testing AI documentation in 2023 and picked Ambience as the only option covering all its specialties with an end-to-end Epic integration. It deployed across 15 specialties and reports, in that vendor case study, 21% more patient face time (40 more minutes per provider per day, from Epic usage data), 24% less time in notes and 18% less after-hours charting. Onvida Health runs Ambience directly inside Epic Hyperspace and Haiku, with chart summaries, ambient notes, coding suggestions and a compliance check on HCC-linked documentation.
Start with Abridge when rollout precedent matters most. Sutter Health expanded Abridge to 1,000+ clinicians within three months of a successful pilot and added patient-facing summaries to the electronic health record; 78% of its clinicians reported better work satisfaction in the vendor case study. CHRISTUS Health ran a 2-month pilot, then went enterprise-wide, and its ambulatory CMIO says implementation took 8 days and that residents are next.
Our recommendation: run both on your three heaviest specialties and ask each vendor how notes and coding suggestions are written back into Epic and who reviews them. Ambience states that its fees are tied to customer outcomes, so ask how that works in a pilot.
Community health centers and small hospitals: Suki, or Nabla if data handling decides
For a health center on athenahealth or MEDITECH, start with Suki; its listed integrations are Epic, Oracle Health, athenahealth and MEDITECH. Utah Navajo Health System's CEO said Suki's integration with the athenahealth EHR was the differentiator for clinicians. CenterPlace Health's chief medical officer said adult health and psychiatry providers saw the most change, with few edits needed afterwards. Rush, an Epic site, rolled Suki out across 28 specialties; its Chief Applications Officer reports that 74% of clinicians said burnout fell in the trial, plus a 10% rise in encounter volumes and an estimated $202 per month more revenue per user.
Consider Nabla when retention and data location are your gating questions. Nabla lists SOC 2 Type II and ISO 27001, with EU data residency for non-US clients. Its trust page says US clients are hosted in US regions, audio is not stored by default, and clinical notes are kept 14 days unless the client changes it. On usedby's Nabla page, a Carle Health vendor case study reports that 55% of clinicians saved at least one hour of documentation time.
Europe and software builders: Tandem for care providers, Corti for developers
For European hospitals and clinics, start with Tandem. Aleris, a Nordic specialist provider, first showed Tandem to stakeholders, validated it with selected specialties, then built a business case and rolled it out across Sweden in phases, with early champions in each specialty and weekly usage reports per clinic. Hôpital Fondation Adolphe de Rothschild tested it with 25 physicians over six weeks, in 1,500+ consultations across nearly 20 specialties, against its DxCare record system. Tandem states that data is not used to train models by default and lists Epic, Cerner, Doctolib and other integrations.
Corti answers a different question. It is an API platform for developers building clinical applications, covering speech-to-text, text generation and ICD-10 and CPT coding, with EU and US data residency. Aleris, Tanner Health and Ramsay Santé appear on its customer pages (OBSERVED), and none of those sources names a task. A hospital without developers should not start here; an EHR or practice software vendor might.
What do US peers suggest for a US reader?
The named US peers split by size. Sutter Health and CHRISTUS Health, both large systems, are CONFIRMED Abridge customers. John Muir Health is a CONFIRMED Ambience customer. Access Health Louisiana, a network of federally qualified health centers, is a CONFIRMED Suki customer named in a 2024 press release alongside other health centers; that source gives no workflow for it individually.
What that suggests: a large system with multi-specialty Epic needs has two precedents to test, Abridge and Ambience. A smaller health center has Suki's health-center cases to compare against. None of this ranks the tools; it tells you whose case studies to read first.
What transfers to a smaller organisation?
Pilot shapes transfer; results do not. Rothschild's physician group, CHRISTUS's two-month pilot and Aleris's specialty champions with weekly usage reports can all be copied at a smaller scale. John Muir's 15-specialty start shows you do not need the whole organisation on day one.
Our recommendation: pick one clinician group per specialty, run a short pilot in your own EHR, and track usage by clinician as Aleris did. Percentages such as Sutter's satisfaction figure, Rush's revenue estimate or John Muir's face-time gain come from other populations and metric definitions, so they cannot be projected onto your clinic.
What the public record does not show
Prices are not published in these sources and are the first question for any vendor. The figures quoted are vendor or customer claims, not independent tests, and no source compares the tools head to head. Sources also do not show clinical accuracy or error rates for any scribe.
Emory Healthcare's Abridge listing names no task, so it supports adoption only. Aleris and Ramsay Santé appear on both Tandem's and Corti's pages, but a shared listing does not show the two are used together.
Method and limits
2026-10-23
Scope: 89 companies classified in Hospitals & Care Providers on usedby with a public line on a Clinical documentation AI tool. Every count on this page covers this sector only, not the tools' whole customer base. Company size is not recorded and is not claimed.
This dated snapshot counts distinct companies in public customer evidence. CONFIRMED means a case study, named testimonial or press source; DECLARED means the company itself told usedby; OBSERVED means a vendor listing; SIGNAL means a job post. All four count, but their strength differs. Removed and unlevelled lines are excluded. Examples are a capped selection verified within 90 days. Vendor facts are sourced records checked within 60 days. Coverage is uneven and missing evidence is unknown. These figures measure neither market share nor product performance; capture dates are not adoption dates.
