Bottom line
ClinicalSim lets clinicians, staff and trainees practice conversations out loud with AI patients in a browser, then scores each conversation against a named communication framework or the institution's own rubric. Every score points to the learner's own words. The vendor calls the scores formative and says it has not completed validation against faculty raters. No price, security certification or integration list is published. It fits a program that wants to pilot communication practice and is ready to ask the vendor for the missing details.
Clinicians, staff and trainees practice spoken conversations with AI patients in the browser; each formative score points to the learner's own words.
No price is published. The FAQ says engagements are scoped and quoted with each institution.
What is ClinicalSim and who sells it?
The site describes it as "the clinical communication intelligence platform for healthcare" (clinicalsim.ai). It names two audiences: health systems, for patient experience, consent and disclosure, and team conversations, and medical education, for residents, fellows and medical students. Use-case pages cover nursing communication training and breaking bad news practice.
The legal entity, location, founding year and funding are not published. The about page names Lauren Rissman, MD, as Chief Medical Officer, board-certified in pediatrics, pediatric critical care medicine, and hospice and palliative care medicine. It also names Vinod Havalad, MD, Gillian Brennan and Jacqueline Ponczek, MD, MS, FAAP, with clinical and education titles, and Ben Conway as co-founder and CEO (clinicalsim.ai/about). The site says "29 academic medical centers and children's hospitals are piloting ClinicalSim" and names none of them.
How does the practice and scoring work?
A learner has a voice conversation with an AI patient designed for the case. Practice is audio only and runs in the browser on a phone, tablet or desktop. The site says it currently supports English (clinicalsim.ai/faq).
Scoring follows the framework a case is built on. The FAQ names SPIKES, the Kalamazoo Essential Elements Communication Checklist, SEGUE, NURSE, REMAP, SBAR, I-PASS, TeamSTEPPS, CANDOR and Calgary-Cambridge, and says programs may add their own rubrics or use the institution's own policy or script. For each scored competency or framework step the platform pulls one or two verbatim excerpts from the transcript. Ratings use a published instrument's own scale where one exists. The UME scale is Demonstrated, Partially demonstrated or Not demonstrated (clinicalsim.ai/methodology).
Cases are reviewed by "practicing physicians with strong academic backgrounds" and run repeatedly before release. A published case's text, rubric and scoring prompts are versioned and locked (clinicalsim.ai/trust). The site does not say which AI model powers the patients or the scoring.
What evidence does the vendor offer?
The vendor publishes no validation or outcome study. The evidence page says "We haven't completed faculty-rater validation on a customer's own rubric," and the FAQ says the company "does not claim that its ratings are more accurate or fairer than faculty judgment." It also says consistency between model runs "isn't agreement with expert human raters." Scores are described as formative and "not intended for employment, credentialing, privileging, licensure, or other high-stakes decisions."
The one data set is a cohort report from an unnamed partner health system: 82 practice conversations by 68 clinicians, with Listen averaging 3.83 and Explain 3.79. The page says examples "don't establish training effectiveness or patient outcomes." Three conference venues are named without titles or dates, and the page adds that a presentation record "isn't evidence of a confirmed product outcome." No paper on the product is indexed in PubMed.
What does the vendor say about privacy and security?
The trust page does not state HIPAA, a business associate agreement, FERPA, SOC 2, encryption, data retention, data residency or single sign-on, and it names no subprocessors or AI model provider. It says voice collection is consent-gated, learners can request erasure and "a learner's recording exists to generate that learner's feedback." Learner data is governed by the agreement with the institution (clinicalsim.ai/trust).
The privacy policy, last updated March 16, 2026, covers contact-form and waitlist data, automatically collected data and cookies. It does not describe voice recordings or transcripts (clinicalsim.ai/privacy). The trust page invites buyers to ask when a security review needs more (clinicalsim.ai/trust).
What does ClinicalSim cost?
"No price is published." The FAQ says engagements are scoped and quoted with each institution. Plans and a free trial are not stated (clinicalsim.ai/faq).
The verdict
ClinicalSim is a communication-practice product with a clear scoring method and traceable feedback, and with no published validation, certifications or price. A training program that wants to pilot spoken practice against SPIKES or its own rubric has a defined product to evaluate. A buyer who needs a security attestation, a named model provider or evidence that scores match faculty ratings has to get those from the vendor before relying on it.
Editorial review last generated November 17, 2026. Synthesized from clinician sentiment, peer-reviewed coverage, and our editorial silo picks. Refined by hand where vendor facts change.





