Weekly digest

Digest Issue #7

The Big Finding, Product Watch, Clinical Reality Check, Head-to-Head, and One Action This Week.

Last updated 2026-11-18

Published 2026-11-18. Sources opened 2026-11-18. Every claim sourced; see the methodology.

One short, sourced email a week: what changed in AI scribe pricing, privacy terms, and features.

The Big Finding

One scribe saved time. The other did not, in the same clinics, over the same weeks. The first randomized trial of ambient AI scribes, published in NEJM AI in November 2025, assigned 238 outpatient physicians across 14 specialties at UCLA to Microsoft Dragon Ambient eXperience (DAX) Copilot, Nabla, or usual care, from November 4, 2024, to January 3, 2025 [1]. Nabla users spent 9.5 percent less time per note than the control group (95 percent confidence interval, 17.2 percent less to 1.8 percent less; P=0.02). DAX users showed a 1.7 percent reduction that was not statistically significant (95 percent confidence interval, 9.4 percent less to 5.9 percent more; P=0.66) [1]. Clinicians used their assigned scribe in about a third of eligible visits: DAX in 33.5 percent of 24,696 visits, Nabla in 29.5 percent of 23,653 visits [1]. Clinicians rated clinically significant inaccuracies as occurring "occasionally" for both products [1].

Three 2026 studies widen the picture, and a new emergency department study extends it further. The largest study yet, published in JAMA in April 2026, followed 8,581 ambulatory clinicians at five US academic health systems, including 1,809 adopters [2]. Adoption was associated with 16.0 fewer minutes of documentation and 13.4 fewer minutes of total EHR time per eight scheduled patient hours, plus 0.49 additional visits per week [2]. EHR time outside work hours did not change significantly [2]. The first large evaluation inside mental health care, published in JMIR Formative Research in March 2026, followed Talkspace's Smart Notes over its first year: 162 full-time and 1,366 contractual mental health providers generated more than 286,000 clinical notes [3]. Weekly use averaged 94 percent among full-time providers and 72 percent among contractual providers for eligible individual therapy sessions [3]. Quality ratings were 97.7 percent thumbs-up among full-time providers and 98.4 percent among contractual providers, but only 6.5 percent of notes carried a formal quality rating, so those percentages describe a small, self-selected slice [3]. Completed sessions rose slightly, from 15.5 to 16.5 per week for full-time providers, with no significant change in total working hours [3]. At Memorial Sloan Kettering Cancer Center, only 190 of 958 clinicians, or 19.8 percent, used the scribe rolled out in summer 2025, and after-hours EHR time fell further among never-users than among adopters, so the drop cannot be credited to the scribe [4].

The newest study, in Annals of Emergency Medicine, brings the same pattern to the emergency department, with a design that answers the measurement problems of earlier work. It followed 275 attending emergency physicians across 20 emergency departments, 3 academic and 17 community, with exposure and outcomes read directly from the EHR and each physician compared with their own unassisted notes [5]. Among 275 attending physicians, 31 percent of encounters were AI-scribed. Scribed notes took 2.4 fewer active minutes after adjustment, with medians of 10.0 versus 12.5 minutes, and ran about 14 percent longer, with medians of 5,220 versus 4,468 characters [5]. Physicians started scribed notes about 15 minutes later, with medians of 31 versus 17 minutes after first seeing the patient. Note completion was not delayed, and 72-hour returns to the emergency department were not increased [5].

Caveats, stated plainly: the trial ran two months at one academic center, was restricted to English-language visits, and time-in-note does not count the minutes spent editing a draft inside the vendor platform, so the saving is measured before the review step [1]. The JAMA study is observational, adoption was voluntary at most sites, which leaves room for selection bias, and the centers were academic [2]. The Talkspace authors state the observational design prevents a causal link, documentation time itself could not be measured on the platform, and all authors were employed by Talkspace [3]. The emergency department study is a controlled retrospective evaluation, not a randomized trial, and it does not name the scribe product [5]. The pattern across all five studies is the same: savings are real and modest, differ by product, depend on use, and leave after-hours work largely unchanged.

Do not map trial results onto scorecard ranks. Nabla is 65/100 on our research scorecard and is not ranked, because no publicly documented BAA was found [7]. Nabla was the product that saved time in the trial. Microsoft Dragon Copilot (formerly DAX) is 59/100, ranked 13th among vendors with a documented BAA on our research scorecard [7]. The trial compared products on time; our scorecard measures published documentation. Those are different kinds of evidence, and our clinical pilot adds a third kind: what the notes actually say, with Heidi at 96.0 and zero critical flags and Mentalyc and Upheal flagged Unsafe as written on two of three encounters each [10].

Product Watch

  • India's public health system picks a multilingual voice scribe. EkaCare was selected by India's National Health Authority as a technology partner to deploy an Ambient AI-enabled Voice Scribe for doctors in public healthcare facilities under the Ayushman Bharat Digital Mission, announced November 12, 2026 [6]. The tool converts consultations into structured clinical notes in Fast Healthcare Interoperability Resources format, supports more than 15 Indian languages including multilingual conversations, and runs on National Health Authority infrastructure powered by Amazon Web Services [6]. Before a note is saved and linked to a patient's Ayushman Bharat Health Account, a verified doctor must review and approve it [6]. The deployment builds on EkaScribe, EkaCare's ambient scribe used by more than 1,000 doctors in private practice [6]. Caveats: this is a government deployment announcement, not a measured outcome. No accuracy, time, or adoption figures have been published for the public-facility rollout.
  • JotPsych drops public prices for two billing products. Our November 12 review update notes that public JotBill and JotAssure dollar figures were removed from JotPsych's pages and now require booking a call, and a Clinical Staff seat was added at $44 per month ($529 per year), while the vendor FAQ on the same page still says $49 per month [8]. JotAudit is now listed as included with done-for-you billing [8]. Core Moonlighting, Basic, and Everything prices, scores, ranks, and verdict are unchanged. JotPsych is 75/100, ranked 4th among vendors with a documented BAA on our research scorecard [7]. A price that moves to a sales call is still a price change: it removes the figure a solo practice uses to compare before buying.

Clinical Reality Check

The language gap is still open, and multilingual scribes are the test case. Last issue's JAMA Network Open finding still stands: same-day note closure gains did not extend to interpreted sessions, and only one of the two tools in that study supported non-English, non-Spanish documentation without an interpreter. EkaCare's 15-plus language support is the kind of capability that could close that gap, but no study has yet shown a multilingual scribe closing it in practice [6]. The claim to check is not how many languages a scribe transcribes. It is whether the clinician can review the draft in the language the session was held in, whether the note stays accurate when the session mixes languages, and whether consent is recorded in the patient's language before anything is captured.

For a therapy practice that works through interpreters, the practical questions before a trial are the same as for any scribe, with one added: run interpreted sessions through the tool, read the drafts for speaker attribution and medication ownership, and check whether the closure benefit the vendor advertises arrives with the interpreter in the room. Until a study measures that, treat multilingual support as a feature to test, not a result to buy on.

Head-to-Head

A $19 bundled scribe against standalone entry prices, with the full stack in view. OptiMantra's AI Scribe, launched November 11, costs $19 per provider per billing cycle for up to 30 notes, or $59 per provider per billing cycle for unlimited notes, drafted inside the chart, covered by the BAA in the OptiMantra agreement, with audio deleted once the transcript is made [11]. The $19 figure is an add-on to the EHR, not a standalone stack. Our psychiatric EHR comparison puts OptiMantra's documented monthly costs for one prescriber at $137, before any scribe is added, and notes that some fees are unpublished [12]. The comparison that matters for a practice choosing an EHR and a scribe together is the documented monthly costs for the full stack: $156 per month with the Limited scribe, $196 with Unlimited, on published figures alone.

Standalone entry prices, all live on our research scorecard, sit in the same range or below it for the scribe alone: Mentalyc Mini at $19.99 per month or $14.99 billed annually for 40 notes, Heidi on a free plan, and Upheal on a permanent free plan with AI Notes at $39 per month [7]. A standalone scribe is cheaper if the EHR is already settled. A bundled scribe is cheaper only when the EHR choice is still open and the bundled stack beats the EHR plus standalone combination. And price is not note quality: our clinical pilot scored Heidi 96.0 with zero critical flags and flagged Mentalyc Unsafe as written on two of three encounters, so the cheapest draft can also be the one that needs the heaviest review [10]. Read the full pricing picture: clinicalaiweekly.com/blog/november-pricing-audit

One Action This Week

Measure your own minutes before you renew or switch. The studies agree on one thing: the saving depends on how much you use the tool and how much you edit. For the next two weeks, log three numbers per clinician: minutes spent reviewing and correcting each draft inside the scribe app, the share of notes that close after hours, and the share of eligible visits actually scribed [1][2][5]. Report adoption per clinician, not per practice, because a practice average hides the clinicians who never use it [4]. Then run the five checks from our pre-signing routine on every draft before you sign: medication names and identity, medication ownership, diagnosis against the examination, only findings observed in the session, and what the note left out [13]. If the minutes do not add up, the problem may be adoption or editing time, not the scribe, and that tells you whether to renew, switch, or stop.

Also new on the site

  • New blog post, November 13: "November Pricing Audit: Twofold's $21 Promo Ends, PMHScribe Drops Annual Prices, and a $19 Bundled Scribe Arrives," the full version of this issue's Head-to-Head, including the annual math on Twofold and PMHScribe and the OptiMantra launch [9].
  • Corrections, November 11, stated plainly: Twofold Health re-scored after a check of official pages supplied by the vendor. Patient instructions or letters re-scored from 0/5 to 5/5, and Workflow from 15/20 to 20/20. Research score 58 to 63. Twofold Health is ranked #9 among vendors with a documented BAA on our research scorecard, previously 13th [7][8]. EHR integration stayed at 0/20, and SOC 2 Type II remains in progress [8]. A clinician-set note-retention window, 42 CFR Part 2 controls, and fuller contract terms were not documented on the pages checked [8].
  • Corrections, November 12, stated plainly: JotPsych review pricing update, as described in this issue's Product Watch. Core prices, scores, ranks, and verdict unchanged [8].
  • Corrections, November 13, stated plainly: November pricing audit blog post published. Scores and ranks unchanged [9].
  • Scorecard unchanged otherwise: 19 tracked vendors, 17 ranked, scores as of 2026-11-11, drawn from 353 linked source pages [7]. Research scores rate public vendor documentation, not note accuracy or clinical outcomes.

Sources

  1. Lukac PJ, Turner W, Vangala S, Chin AT, Khalili J, Shih YT, Sarkisian C, Cheng EM, Mafi JN, "Ambient AI Scribes in Clinical Practice: A Randomized Trial," NEJM AI, 2025;2(12):AIoa2501000, https://pmc.ncbi.nlm.nih.gov/articles/PMC12768499/ (opened 2026-11-18)
  2. Rotenstein LS, Holmgren AJ, Thombley R, et al., "Changes in Clinician Time Expenditure and Visit Quantity With Adoption of Artificial Intelligence-Powered Scribes: A Multisite Study," JAMA, 2026;335(16):1408-1417, https://doi.org/10.1001/jama.2026.2253 (opened 2026-11-18)
  3. McCrudden KE, Swirbul MS, Peake EE, Rodio MJ, Padmanabhan A, "AI-Powered Documentation for Mental Health Providers: Retrospective Observational Mixed Methods Study," JMIR Formative Research, 2026;10:e84628, https://formative.jmir.org/2026/1/e84628 (opened 2026-11-18)
  4. Healio, "Oncologists give ambient AI scribes positive reviews," October 27, 2026, https://www.healio.com/news/hematology-oncology/20261027/an-exciting-time-oncologists-give-ambient-ai-scribes-positive-reviews (opened 2026-11-18), reporting Baldwin-Medsker AE, et al., "Ambient Artificial Intelligent Scribes in Oncology: Impact on Clinician Experience and Documentation Burden," JCO Oncology Practice, 2026, https://doi.org/10.1200/OP-26-00822
  5. Martini WA, Heaton HA, Walker L, Urumov A, Rappaport DE, Hodgson NR, "Ambient Artificial Intelligence Scribe Use and Objective Documentation Effort Among Emergency Physicians," Annals of Emergency Medicine, https://doi.org/10.1016/j.annemergmed.2026.08.737 (opened 2026-11-18)
  6. EkaCare, "EkaCare Selected as a Technology Partner for National Health Authority's Ambient AI-enabled Voice Scribe Solution for Doctors" (NewsVoir release), https://www.newsvoir.com/release/ekacare-selected-as-a-technology-partner-for-national-health-authority-apos-s-ambient-ai-enabled-voice-scribe-solution-for-doctors-37383.html (opened 2026-11-18)
  7. ClinicalAIWeekly, "AI scribe research scorecard" (19 tracked, 17 ranked; scores as of 2026-11-11; 353 linked source pages), https://www.clinicalaiweekly.com/scorecard (opened 2026-11-18)
  8. ClinicalAIWeekly, "JotPsych review" (November 12 pricing update), https://www.clinicalaiweekly.com/reviews/jotpsych (opened 2026-11-18); "Twofold Health review" (November 11 rescore), https://www.clinicalaiweekly.com/reviews/twofold (opened 2026-11-18)
  9. ClinicalAIWeekly, "November Pricing Audit: Twofold's $21 Promo Ends, PMHScribe Drops Annual Prices, and a $19 Bundled Scribe Arrives" (blog, published 2026-11-13), https://www.clinicalaiweekly.com/blog/november-pricing-audit (opened 2026-11-18)
  10. ClinicalAIWeekly, "Psychiatric Note Accuracy Pilot: Six Vendors Scored" (Heidi 96.0 with 0 critical flags; Mentalyc and Upheal flagged Unsafe as written on two of three encounters; not a ranking), https://www.clinicalaiweekly.com/clinical-benchmarks/results (opened 2026-11-18)
  11. OptiMantra, "AI Scribe" product page (Limited $19 / Unlimited $59 per provider per billing cycle; covered by the BAA in the OptiMantra agreement; audio deleted once the transcript is made), https://www.optimantra.com/ai-scribe (opened 2026-11-18); OptiMantra, "OptiMantra Launches AI Scribe, Bringing Native AI Charting to Independent Practices" (PR Newswire, November 11, 2026), https://www.prnewswire.com/news-releases/optimantra-launches-ai-scribe-bringing-native-ai-charting-to-independent-practices-302936679.html (opened 2026-11-18)
  12. ClinicalAIWeekly, "Psychiatric EHR comparison 2026" (OptiMantra documented monthly costs for one prescriber $137, some fees unpublished), https://www.clinicalaiweekly.com/practice-stack/ehr-comparison/ (opened 2026-11-18)
  13. ClinicalAIWeekly, "The Pre-Signing Routine: Five Checks Before You Sign an AI-Scribed Note" (blog, published 2026-11-09), https://www.clinicalaiweekly.com/blog/ai-scribe-pre-signing-routine (opened 2026-11-18)

Who curates this site

Curated by a practicing, board-certified psychiatrist, not named to avoid conflicts. From the site data: 19 vendors tracked, 354 distinct external source pages linked. Last site update 2026-11-20; last targeted vendor verification 2026-11-12; last complete research audit 2026-11-04. How we work · Editorial policy.

Get The Behavioral Health AI Brief

One short, sourced email a week: what changed in AI scribe pricing, privacy terms, and features. Free, unsubscribe anytime.