AUDOVA™ listens to a recorded clinical encounter and returns structured, evidence-based coaching across four dimensions — in under 60 seconds, in English and Spanish. Coaching, not diagnosis: it is not a medical device. The AI suggests; the clinician always decides.
To help every clinician deliver more human, safer, and more accurate care — putting AI in service of clinical judgment and patient wellbeing.
We turn each visit into a chance to learn, translating modern medicine into clear, evidence-based, actionable feedback — so improvement becomes a sustainable habit, not a matter of chance. The AI suggests; the human decides.
To define a new, AI-augmented medicine — where technology amplifies the most human part of caring, held to the highest ethical, professional, and humanitarian standards.
A future where no clinician practices alone, but alongside a rigorous, respectful coach that helps them grow across their whole career — with care that is excellent in any language and any setting.
Because communication is part of the treatment.
What if the most powerful tool in medicine isn't a scalpel, a scan, or a prescription — but a sentence?
As a primary care physician, I became obsessed with something medical school barely touched: communication. Not just the words I chose, but the pauses between them. The posture. The eye contact. The micro-expressions that told a patient, "I'm here. I'm listening. You matter."
I spent years studying what made healing encounters actually heal — refining my approach, observing what connected and what fell flat, drawing on my background in neuroscience, research, teaching, and clinical leadership. And I discovered something uncomfortable.
Physicians and advanced practice providers spend years mastering anatomy, pharmacology, and diagnostics. But the art of communication? Motivational interviewing? Building genuine rapport? Those get a footnote at best.
As a medical director supervising other providers, I saw the gap up close every day. Brilliant clinicians struggling to connect. Patients leaving visits feeling unheard. Diagnoses missed — not from lack of knowledge, but from lack of listening.
I knew this wasn't a character flaw. It was a training gap. And I wanted to build something to close it.
A great teacher once told me: put three stones in your pocket before every meeting. Each time you speak, take one out. When they're gone, you're done. You have three chances to say something — make them count.
That idea stuck with me. It reshaped how I practiced medicine, and it became part of the DNA of what we built. Three stones became a symbol for AUDOVA™ — a reminder that in every encounter, listening is the real superpower.
Three chances to speak. Infinite chances to listen. The best diagnoses — and the deepest trust — live in the silence between the stones.
AUDOVA™ is built on the conviction that voice and listening are clinical tools. We help providers find the right words. Not scripted phrases or canned empathy, but authentic, effective communication shaped by real-time feedback and the science of human connection.
AUDOVA™ isn't a checklist. It's a purpose-built AI that analyzes patient interactions and gives you actionable feedback across the dimensions that matter most:
Are you guiding change, or directing it?
Does the patient feel understood — not just heard?
Are you building trust that lasts beyond the visit?
Is the patient a partner in their own care?
Are you translating expertise into understanding?
The more you use it, the more you grow. AUDOVA™ learns with you, surfaces patterns you can't see yourself, and turns every patient encounter into an opportunity to get better at the thing that matters most.
We're building toward CME accreditation — so the time you spend improving your communication also counts toward your continuing education requirements. Learn by doing, grow by listening, and earn credit along the way.
More to come.
Clinicians get thousands of hours on pharmacology and diagnosis — and a handful on how to actually talk with patients.
Communication failure is a leading root cause of sentinel events.
The Joint Commission
In losses across 1,744 cases where communication broke down — roughly 30% of malpractice claims in a 5-year study.
CRICO Strategies
Of the information shared in a visit is accurately recalled by the patient afterward.
Kessels, 2003
The same recording, read from four useful angles. Each pillar uses recognized clinical references — not ad-hoc scoring.
Motivational interviewing, empathy, and shared decision-making — scored with validated instruments, not ad-hoc rubrics.
MITI 4.2.1 · Four Habits
Surfaces knowledge gaps that appear inside the visit and maps them to targeted CME.
ADA · ACC/AHA · USPSTF
Flags preventive services and social drivers of health that went unaddressed in the conversation.
USPSTF · CDC · SDOH / Z-codes
Identifies quality measure and documentation opportunities grounded in the encounter itself.
MIPS · HCC · HEDIS
Every output is evidence-grounded and passes an explicit anti-hallucination check before the clinician ever sees it.
The clinician records the visit or pastes a transcript. Audio is transcribed with speaker separation; language, visit type, and context are detected.
A layered AI pipeline reads the encounter across the four pillars, grounded in verified clinical knowledge.
The clinician receives a coaching report: scores by pillar, phrases to improve, screening alerts, and a 3-step action plan.
Under 60 seconds, end to end · English & Spanish · audio minimized, access-isolated, deletable
Why the feedback is trustworthy: every layer is grounded and checked.
Reads clinical entities, preventive services, and conversation dynamics from the transcript.
Each pillar is evaluated with its validated instrument and grounded against verified clinical knowledge.
Every clinical statement is checked against the transcript and its sources. If it isn't grounded, it is removed.
A single, coaching-toned report — never punitive, always actionable.
Source-grounded · traceable · human-in-the-loop
Motivational Interviewing Treatment Integrity: a fidelity yardstick for how MI is actually practiced. The first two steer the conversation, the last two build the relationship — every encounter scored the same way.
Steering the conversation
Do you draw out the patient's own reasons for changing? Get them saying why — not you telling them.
Do you avoid feeding the arguments for staying the same? Redirect toward change, don't reinforce the status quo.
The quality of the relationship
Do you treat the patient as the expert and share the decision? Power-sharing, not questions without follow-through.
Do you truly understand what they mean — even what they haven't said? Not sympathy or warmth: accurate understanding.
We don't invent scores. Communication is assessed with MITI 4.2.1 and the Four Habits Coding Scheme — instruments used in clinical research.
Grounded in ICD-10-CM, LOINC, RxNorm / SNOMED, USPSTF, CDC, MIPS, HCC, and HEDIS — over 440,000 curated references.
If a statement can't be traced to the encounter and a source, it never reaches the clinician. "If not certain, say so" is enforced in the pipeline itself.
We never train or fine-tune any model on patient or physician data.
Every output cites the verified knowledge that informed it.
Outputs are prompts to reflect, not directives. The physician decides.
A physician's data is never shared with employers or payers without explicit written consent.
Encrypted in transit and at rest, row-level isolation, audio minimized and deletable.
Bilingual with per-language calibration and a program to measure and reduce any gap.
The legal baseline for PHI — minimum-necessary use, the Business Associate chain, safeguards, and breach response.
Responsible AI across the pipeline — accuracy, anti-hallucination, source-grounding, bias management, and human oversight.
Cybersecurity — access control, encryption, secrets management, monitoring, and incident response.
Privacy across the data lifecycle — minimization, consent, transparency, and individual rights.
HIPAA is the baseline, not the goal — a single unified governance spine, with versioned, auditable documents behind every commitment.
"Three stones in the pocket before each encounter — three chances to speak, and infinite chances to listen."
AUDOVA™ is currently in BETA testing with a validated model. HIPAA safeguards are in place, and we can run real patient encounters. We improve clinical communication with patients and are open to pilots in your organization.
The ask
A 20–30 minute live demo
We run a sample encounter end to end, show the exact coaching report a clinician would receive, and explore whether a small, simulated-data pilot could fit your organization.
Request a demo