AI medical scribe startups raised $1.56B across 17 deals in 2025. This amount exceeds last year’s $350M by more than four times. Abridge raised $550M in 2025, which gave the company a $5.3B valuation. Abridge investors included Andreessen, Elad Gil, IVP, Redpoint, Bessemer, CVS Health, Kaiser, and Mayo Clinic. Ambience Healthcare secured $243M in 2025, pushing its total funding to $345M. Ambience investors included Oak HC/FT, Andreessen Horowitz, OpenAI Startup Fund, Kleiner Perkins, and Optum Ventures. Heidi secured $340M through a $100M Series C and a $240M growth investment. The company reports that its software handled 2.8 million notes each week across 190 countries. Heidi has seen use in 20 million patient interactions since February 2024. The company reached a valuation of A$1.26 billion. Commure reached a $6B valuation following a merger between Athelas and Augmedix. Suki AI has raised more than $165M in total funding.
Clinician burnout and time savings
Clinicians spend 13 hours per week on tasks like documentation, order entry, and test result reviews. For every 8 hours of scheduled patient time, physicians spend more than 5 hours in the EHR. A 2025 Tebra survey found that documentation causes burnout for 16% of all providers and 26% of primary care physicians. Over 20% of physicians spend more than 8 hours per week on EHR work after normal hours. You already know that administrative work consumes the majority of a physician’s workweek. The 2025 Yale-led study published in JAMA Network Open tracked 263 doctors and advanced practice providers across six different U.S. health systems for a full 30 day period to measure how Abridge affected clinician burnout. This study showed burnout dropped from 51.9% to 38.8% within 30 days. At Mass General Brigham, clinicians saw a 21.2% drop in burnout after 84 days using ambient AI. At Emory Healthcare, clinicians saw a 30.7% improvement in documentation-related wellbeing after 60 days. The Permanente Medical Group reported that its physicians saved 15,791 hours of documentation time in a single year. High users of AI scribes saved two and a half times more time per note than low users. The Permanente evaluation of 7,260 physicians across 2.5 million patient encounters showed that clinicians who used the technology most frequently experienced the greatest time savings effect. Eighty-four percent of physicians said the technology had a positive effect on patient interactions. Forty-seven percent of patients said their doctor spent less time looking at their computer during a visit, and 39% said their doctor spent more time speaking directly to them.
Pricing and legal risks
| Vendor | Estimated Monthly Cost per Provider | Primary Integration |
|---|---|---|
| Nuance DAX | $369 – $830+ | Epic and Oracle Health |
| DeepScribe | ~$750 | Various (e.g., Athenahealth, NextGen) |
| Abridge | ~$208 | Epic |
| Vero | $69+ | Manual export/copy |
Nuance DAX maintains deep workflows inside Epic and Oracle Health. Microsoft acquired Nuance in 2022 for $19.7 billion. The software uses a training corpus of clinical speech and documentation collected over decades. This provides specialty-specific accuracy for cardiology, orthopedic, and psychiatry terminology. However, smaller practices find the Nuance enterprise procurement process difficult. Abridge provides native Epic integration via Haiku and Hyperdrive. Abridge also produces patient-facing summaries at an eighth-grade reading level. DeepScribe earned a 98.8 out of 100 score in the January 2025 KLAS Research Emerging Company Spotlight. It received A+ grades in Product, Value, Operations, Relationship, Loyalty, and Culture. DeepScribe uses specialty-trained AI models for oncology, cardiology, orthopedics, urology, gastroenterology, and neurology. Roughly 40% of all cancer care visits in the U.S. use the DeepScribe platform. Vero offers a lower price point but requires manual copy or export after review.
Liability remains a significant concern for providers. Providers bear the legal responsibility for any AI recommendation, whether clinical or documentation-based. In Texas, the law requires providers to document whether they used AI to record patient encounters. North Carolina also has legislation regarding AI documentation. Improper use of AI can lead to up-coding errors or higher billing codes than performed, which remains the provider’s responsibility. A proposed class action in November 2025 against Sharp HealthCare alleged that an ambient AI scribe recorded 100,000 visits without proper consent. Eighty-five percent of healthcare AI investment goes to startups without proven compliance track records. The decision in 2026 rests on whether a practice needs high-end enterprise integration or simple, low-cost documentation support. Will regulators eventually mandate specific transparency standards for every AI-generated medical note?
