Yardstick Research tear-sheet / healthcare clinical cohort

Methodology · how we score · rubric weights in plain sight · vendors received this sheet seven days before publication and could flag factual errors, never rankings

Deepscribe

Identity

• Founded: 2017 by Akilesh Bapu, Matthew Ko, and Kairui Zeng (out of UC Berkeley AI Research Lab). [VENDOR-CLAIMED - https://www.deepscribe.ai/about]
• HQ: San Francisco, California. [THIRD-PARTY - https://www.crunchbase.com/organization/deepscribe]
• Legal entity: DeepScribe, Inc.
• Funding: ~$60-61M total raised. Seed $5.2M (May 2021), Series A $30M (Jan 2022, Index Ventures lead; Scale AI CEO Alex Wang + Figma CEO Dylan Field participated), Series B $24M (Oct 2023). No public Series C as of May 2026. [THIRD-PARTY - https://www.crunchbase.com/organization/deepscribe, https://www.businesswire.com/news/home/20220111005911/en/DeepScribe-Raises-%2430M-To-Become-First-Widely-Accepted-Application-of-Voice-AI-in-Healthcare]
• Headcount: [UNKNOWN] - Glassdoor profile aggregates 165 reviews; no point-in-time figure published.
• Recent news (last 12 months):
• Dec 2024 - Ochsner Health selects DeepScribe for 4,700 clinicians. [VENDOR-CLAIMED - https://www.deepscribe.ai/resources/ochsner-health-selects-deepscribe-to-bring-ambient-ai-to-their-4-700-clinicians]
• Sep 2025 - Oncology study + accelerated growth: 80%+ adoption among oncology orgs DeepScribe serves, on track for 3.1M cancer-care visits/yr, monthly oncology visits quadrupled YoY. New named oncology customers: USON, AON, OneOncology, OnCare, SCRI, CARTI. [VENDOR-CLAIMED - https://www.prnewswire.com/news-releases/deepscribe-solidifies-ambient-ai-leadership-in-oncology-with-new-study-and-accelerated-growth-302557045.html]
• Early 2025 - KLAS Emerging Company Spotlight: 98.8/100 overall, A+ across all six categories. [VENDOR-CLAIMED + THIRD-PARTY - https://www.deepscribe.ai/, https://www.ehrsource.com/articles/ambient-ai-scribes-comparison/]
• March 2024 - Real-time AI-generated notes unveiled at HIMSS. [VENDOR-CLAIMED - https://hitconsultant.net/2024/03/12/deepscribe-unveils-real-time-ai-generated-clinical-notes/]
• Archetype: Specialty-tuned ambient AI medical scribe + AI Coding + pre-visit intelligence ("SmartPrep") + customization studio. Marketed as the "Ambient Operating System." Distinct positioning: oncology + complex specialty depth (cardiology, GI, neurology, urology) rather than general primary care or broad inpatient CDI.

Total score: 78.75 / 100

• Stage fit:
• Foundation (<40 readiness): no - no free tier, no self-serve trial, sales-led only.
• Pilot (40-59): conditional - only for groups large enough to negotiate a structured pilot. Single-clinic ambulatory practices may not clear vendor attention threshold.
• Scale (60-79): yes - Ochsner 4,700-clinician deployment, USON/AON/OneOncology oncology footprint, 2,800+ clinics & health systems served are production-scale signals.
• Optimization (80+): yes - AI Coding tightly coupled to note workflow + SmartPrep pre-visit intelligence + Customization Studio is the structural fit for an optimization-stage practice measuring against documentation burden + revenue-integrity.
• One-line verdict: Specialty-tuned ambient scribe with strongest oncology footprint in the cohort and a KLAS 98.8 spotlight score, capped below cohort leaders by the absence of any DeepScribe-specific peer-reviewed validation, no published hallucination rate, no Cerner/Meditech integration, and a Series B last round that lags peer fundraising by an order of magnitude.

Headline numbers

Dimension scores

Pricing detail

No public pricing page. All contracts are sales-led. Bands below are [THIRD-PARTY ESTIMATE] from independent industry analyses cross-referenced where possible.

• No-EHR-integration tier: ~$400/provider/month (~$4,800/year). [THIRD-PARTY ESTIMATE - https://softwarefinder.com/emr-software/deepscribe/pricing]
• With-EHR-integration tier: ~$500/provider/month (~$6,000/year). [THIRD-PARTY ESTIMATE - https://www.healos.ai/blog/deepscribe-pricing-features-cost-and-the-best-alternatives-in-2025]
• Full-suite high band: up to ~$750/provider/month (~$9,000/year) depending on usage and scale. [THIRD-PARTY ESTIMATE - https://softwarefinder.com/emr-software/deepscribe/pricing]
• Small-practice example: ~$2,500/month for 10 users. [THIRD-PARTY ESTIMATE]
• Enterprise example: ~$20,000/month for 100 users (negotiated). [THIRD-PARTY ESTIMATE]
• Implementation fee: [UNKNOWN]
• Volume + multi-year discounts: [UNKNOWN] No published schedule.

Modeled annual cost (Yardstick estimate, $500/provider/month midpoint, before volume discount):

• 25 clinicians: ~$150K/year
• 250 clinicians: ~$1.5M/year
• 2,500 clinicians: ~$15M/year

Sits roughly between Ambience (~$2,800-$5,000/clinician/yr base-to-suite, sales-led) and Suki ($299-$399/provider/month) in the cohort.

Integrations

• EHR (named on vendor site): Epic, Athenahealth, eClinicalWorks. [VENDOR-CLAIMED - https://www.deepscribe.ai/]
• Oncology systems: Flatiron, Ontada, MM20, OMS. [VENDOR-CLAIMED - https://www.deepscribe.ai/]
• NOT named: Cerner / Oracle Health, Meditech, NextGen, Allscripts/Veradigm. [UNKNOWN - likely none integrated as of 2026-05-23]
• Underlying-model providers: Proprietary HealAI (primary); OpenAI GPT-4 (adjunct via prompt engineering). [VENDOR-CLAIMED - https://www.deepscribe.ai/resources/deepscribe-outperforms-gpt-4-by-59-percent-on-ai-medical-scribing]
• NOT documented: Epic App Orchard / Toolbox / Hyperdrive / Haiku embed level; Athenahealth Marketplace certification; published FHIR resource list; SMART on FHIR conformance; HIE / TEFCA QHIN participation; RCM platform connectors.

Integration penalty verdict: SOFT for Epic / Athena / eClinicalWorks buyers and for oncology-specific systems (Flatiron / Ontada). HARD for Cerner / Oracle Health, Meditech, NextGen, Allscripts/Veradigm buyers - DeepScribe is not deployable as a chart-integrated tool in those environments and must be marked excluded with the integration-mismatch reason rendered visibly per Yardstick methodology.

Scale penalty verdict: NONE. Ochsner 4,700-clinician deployment + USON/AON/OneOncology footprint + 2,800+ clinics & health systems served + ~5M oncology visits/year are production-scale signals across both enterprise and group-practice buyer sizes.

Editorial assessment

DeepScribe is the specialty-care-first ambient AI scribe in the healthcare clinical cohort. Where Ambience plays for enterprise IDN bake-offs and Abridge plays for KLAS volume + Epic-Kaiser scale, DeepScribe plays for oncology depth and complex-specialty tuning. The USON / AON / OneOncology / SCRI / CARTI roster plus ~3.1M projected annual cancer-care visits puts DeepScribe at the top of the cohort for oncology buyers. Ochsner Health's 4,700-clinician selection is the strongest enterprise validation point on file. KLAS 98.8 Spotlight + 99.92% vendor-reported note satisfaction + Real-Time generation + SmartPrep pre-visit + AI Coding gives DeepScribe a workflow-surface story at parity with cohort leaders on clinician experience.

Three bounded weak spots. First, no DeepScribe-specific peer-reviewed validation study - the cohort has them at the top (Nabla in JAMA Network Open via UCLA, UW Health/Abridge in NEJM AI, DAX in multiple journals) and DeepScribe leans on internal benchmarks (DeepScore, "+59% vs GPT-4"). A buyer's clinical-safety committee will note this. Second, no published hallucination rate. Third, EHR breadth is solid for ambulatory but absent on Cerner / Oracle Health, and the Epic embed level (Toolbox / Hyperdrive / Haiku) is not documented. Hospital InfoSec teams used to Ambience-level disclosure may flag the gap.

Funding-stage signal is the other watch item. Last public round was October 2023 ($24M Series B); peers raised an order of magnitude more in the same window (Ambience $243M Series C July 2025, Abridge $250M+ across 2024-2025). Either DeepScribe is capital-disciplined or undercapitalized - public data does not distinguish.

For an oncology network, a specialty group practice in cardiology / GI / neurology / urology, or an ambulatory health system on Epic, Athena, or eClinicalWorks that values specialty-tuned scribing + coding-integrated note generation, DeepScribe is on the short list. For an inpatient-CDI-heavy enterprise IDN, an academic medical center demanding peer-reviewed validation, or a Cerner / Meditech / NextGen / Allscripts shop, the cohort offers stronger options today.

Best for

• Stage: Scale and Optimization. Pilot is conditional. Foundation is not a fit.
• Company profile: US specialty group practices and ambulatory health systems on Epic, Athena, or eClinicalWorks. Oncology networks (community + academic-affiliated) are the strongest fit segment. 25-5,000 clinicians.
• Industry fit within healthcare: Strongest for oncology, cardiology, GI, neurology, urology, orthopedics. Conditional for primary-care-only practices (peers compete harder there). Not a fit for inpatient-CDI-led enterprise deployments without a documented CDI surface.
• Sales motion: Sales-led, multi-month procurement; pilot patterns vary by buyer size.
• Annual tooling budget: $150K (25 clinicians) to $15M+ (2,500 clinicians) before negotiated discounts.
• Skip if: You run (a) Cerner / Oracle Health, Meditech, NextGen, or Allscripts/Veradigm as your EHR - DeepScribe does not integrate. (b) Need a vendor-published hallucination rate with methodology pre-contract. (c) Need a DeepScribe-specific peer-reviewed validation study. (d) Need HITRUST CSF certification confirmed pre-contract. (e) Are an early-stage AI adopter looking for self-serve trial - DeepScribe is sales-led only.

Right-of-reply gaps

The following items are flagged [UNKNOWN] in the dossier and should be put to DeepScribe during the factcheck pipeline:

• Peer-reviewed validation study - any in submission, preprint, or planned?
• Hallucination + omission rate with methodology (sample, specialty mix, scoring rubric).
• HealAI vs GPT-4 routing logic - which workflows route where, what BAA flow covers OpenAI sub-processor?
• Epic embed level - Toolbox / Hyperdrive / Haiku?
• Cerner / Oracle Health, Meditech, NextGen, Allscripts integration roadmap?
• HITRUST CSF status - held, in progress, or not planned?
• SOC 2 Type I vs Type II + most recent audit date.
• Enumerated sub-processor list.
• Contract default on PHI-for-model-training.
• Implementation SLA - first-clinician-live timeline for a 250-clinician deployment.
• Current ARR + headcount.
• Series C status - planned, in progress, deferred?
• MedQA / MedHELM / MedBench scores for HealAI.

Sources

See full source list in the research dossier: content/vendor-research/deepscribe.md