The Two-Sided Clinical Intelligence Platform

Aurevia brings the patient's between-visit longitudinal story into every physician decision - and closes the loop after discharge.

Pre-Seed · $1.5M SAFE · $10M Cap

$1.5M

Pre-Seed SAFE

$10M

Post-Money Cap

2

Signed Evaluative Pilots

4

Data Moats

The Problem: Snapshot vs. Story

The ED sees a snapshot. The patient lived the film.

Emergency physicians make life-changing decisions in minutes using only what is in front of them. The months of patient history that would change everything are not in the room.

  1. Physicians operate without the between-visit longitudinal story - The chart holds this visit, it does not hold their heart rate spike, their family conversations and the missed medicine doses at home.
  1. Documentation happens after the decision - Severity goes unproven, comorbidities uncaptured. Focus on optimizing the clinical decision, documentation as byproduct.
  1. Discharge ends the loop - What happens after the patient leaves is invisible. Re-admissions accumulate. We help patients navigate the uncertain post-visit terrains.

5,000 waking hours a year uncaptured

The vast majority of a patient's life events, symptoms, and behaviors outside of clinical visits remain unknown.

Ten-minute visits where AI has focused

AI solutions have traditionally optimized for short, episodic encounters, missing the broader context.

■ ~30 min/year captured by AI    ■ 5,000+ hours uncaptured — Every square represents a week of a patient's life. Only a handful of moments ever reach the clinical record.

Ammar Ahmed, DO
Founder & CEO

It began with my mother.

Skin, joints, heart. Every doctor, primary care and ER, had ten or fifteen minutes and one fragment. Two years of symptoms and fear before one physician spent two hours with her and found lupus. Shortly after, another physician told her she had seven years to live.

– AMMAR AHMED, DO · FOUNDER AND CEO

I have been on both sides of this: as her son, and as the physician with ten minutes and a chart scattered across systems. Not the doctors' fault. Structural. Every one of her physicians was good. None had time to know her. We make life-changing decisions under intense time and cognitive constraints, from a summary the patient gives in ten minutes of a story that took years.

Not the doctors' fault. Structural.

Every one of her physicians was good. None had time to know her. The problem is not intelligence or compassion - it is a system that gives physicians fragments and calls it care.

The Patient Side

Patients suffer for years before care is optimized, if ever, and start to believe their doctors are not listening. It was never an intelligence problem.

Two Sides, One Puzzle

An AI that serves only the physician or only the patient leaves the puzzle unsolved. Connecting the two sides is the entire premise, and it is my life's work.

The Solution: One Record. One Decision. One Loop That Closes.

BEFORE & DURING

  • Patient symptoms, medications, outside records
  • Voice notes, images, wearable data
  • Consent-gated; patient controls what flows

AT THE DECISION

  • Every new event filtered for relevance
  • Only events that change a decision are surfaced
  • Surfaces changes to diagnosis, treatment, or disposition; silent otherwise

AFTER DISCHARGE

  • Follows the patient post-visit
  • Resolves uncertainty, provides guidance
  • Resolves meaningful clinical deterioration
  • Outcome connected back to original decision
  • Identifies and automates patient care bottlenecks.

Integration & Deployment

  • Integration: Epic · Apple Health · Google Health Connect · Oura
  • Deployed: 1 outpatient site live · ~1,000 patient cases analyzed longitudinally over two years

The Delta: Same Doctor. Same AI. Different Data.

Resolving uncertainty improves throughput, & clinical accuracy

Patient: Walter Reyes, 78 - EMS arrival, short of breath for three days.

❌ WITHOUT THE BETWEEN-VISIT STORY

5 Differential Diagnoses

  • Pneumonia
  • ACS
  • COPD exacerbation
  • Pulmonary embolism
  • HF exacerbation (ranked #5)

Extensive Orders

  • CXR, broad labs, troponins
  • CTA considered
  • Empiric antibiotics
  • Admission

Timeline

  • Diagnosis reached at hour five
  • Six hours from door to admission

✅ WITH THE BETWEEN-VISIT STORY

Diagnosis: Acute on Chronic HF

  • Ranked #1 immediately (moved from #5)
  • Home scale weight gain confirmed
  • Daughter's swelling observations
  • Diary entries + outside EF record

Targeted Order Bundle

  • IV furosemide 80mg
  • BNP, BMP
  • Continuous telemetry
  • Daily weights, strict I&Os

Outcomes

  • Diuresis begun in first hour
  • Treated at door, admitted with proof
  • Shorter throughput
5→1

Differential Rank

HF moved from #5 to #1 immediately

5hrs

Time Saved

Diagnosis at hour one vs. hour five

5hrs

Time in ED Reduced

Five hours of ED stay avoided

1,840

BNP Confirmed

Lab results validated Aurevia's prediction

The Four Data Moats: Why Aurevia Compounds.

Model capability is no longer differentiating. Our aim is the proprietary data The model operates on. A data corpus that compounds with each deployment.

Frontier models have little training data on what happens inside a clinical workflow. Ours grows with every deployment, and every deployment makes the next case better.

1

LONGITUDINAL PATIENT CONTEXT

Explains the case - The months between visits: symptoms, medications, outside records, voice, images, wearable data, in the patient's own words, with timestamps. In no EHR. In no training corpus.

2

PHYSICIAN REASONING TRACES

Explains the decision - Step-by-step records of how expert physicians work a case: what they notice first, what they rule out and why, when they judge it safe to act. Accumulate in real clinical workflows

3

OUTCOMES LINKED TO DECISIONS

Explains what the decision caused - A clinical action linked to what the patient reports on day four and what lands in the record after. Closes the discharge gap. Monitors post-discharge trajectory. Catches drift and harm early. Labels the decision with what followed.

4

WORKFLOW BEHAVIOR DATA

Explains how to deliver the next insight - What physicians open, adopt, and set aside on shift, and how the interface is shaped by it. Adoption data no model holds. It trains the silence filter. The experiment is already run.

The Aurevia Pipeline

Story → Chart → Proactive event-driven reasoning → Decision → Outcome

Five stages. One continuous loop. Every outcome compounds the model.

01 · PATIENT RECORD

  • Text and voice notes
  • Photos and outside PDFs
  • Wearables and CGM
  • Home devices

BETWEEN VISITS, CONTINUOUS

02 · EHR + CONSENT

  • Years of history
  • Labs, vitals, imaging
  • Ambient voice on/off
  • Live encounter

PATIENT CONTROLS

03 · AUREVIA FILTERS

  • Every new event, weighed
  • Kept only if it changes a decision
  • Structured and put in context
  • Silent otherwise

CHART DIGGING & FILTERING, SO THE DOCTOR DOESN'T HAVE TO

04 · PHYSICIAN DECIDES

  • Accept or reject
  • A word or a click
  • Differential updated
  • Orders written

HUMAN IN THE LOOP

05 · OUTCOME CLOSES

  • Post-discharge patient report
  • Readmission or discharge
  • Links back to the decision
  • Compounds the model
  • Trajectory changed

LOOP CLOSED

"The clinical uncertainty lives in the lost information of patients' daily lives. Labs, vitals, imaging, and what the physician says arrive the same way. Aurevia weighs each new event as it lands and shows the physician only what changes the next decision - so the sorting and the judgment on what matters happen before the physician ever sees it.

The Patient Side: Between Visits

A health record the patient owns, and a proactive agent that walks them to the next visit.

SET UP BY A QR CODE ON THE DISCHARGE PAPERS. CONNECT WHAT YOU ALREADY HAVE, ONCE.

🔶 THE RECORD

Ingests text, voice, photos, wearables, CGM, cuffs and scales, outside records, pharmacy fills. One consent; information flows. It summarizes, filters and organizes, so the physician sees the few facts that changed the picture, never a raw stream.

🔶 A PROACTIVE AGENT MONITORS THE PATIENT'S TRAJECTORY

Always, and especially after discharge, watching for improvement or decline in the same streams the record ingests. It steps in when a number moves: weight up two kilos in three days, did anything change, did you miss a pill? One question, answered by voice. Silent otherwise.

🔶 GUIDANCE

Helps the patient through the clinical uncertainty and anxiety between visits: what do I do about my pain in the procedure site, should I go to urgent care today, the PCP this week, or stay home and watch. The patient stops second-guessing.

🔶 NAVIGATION AND AUTOMATION

Identifies and engages the right people and does the work: finds the relevant caregiver or doctor, books the appointment, updates prescriptions, coordinates care across them, and flags insurance gaps before they become a bill.

🔶 THE SYNC

The PCP or specialist sees it inside their own workflow, is aware, and can ask a question back. Both sides work from the full clinical picture and know the next step.

Market Opportunity: $4B TAM, Both Ends Still Open.

~$4B

TOTAL ADDRESSABLE MARKET

(Hospitals with EDs plus US primary care)

~$1.6B

SERVICEABLE MARKET

(3,100 IPPS hospitals, where the DRG mechanism is proven today)

~$3M

ARR 12 TO 15 MONTHS

(All four systems converted at $750K; then enterprise rollout across their 27 EDs, ~$20M, in 24 to 30 months)

Why Now

1. AI can now collect, synthesize and filter continuous patient data.

Until now, between-visit data was a raw dump no physician had time to read, so it was never captured.

2. Frontier models are generalizing and winning;

everyone utilizes the same models, including our competitors. Model capability is no longer a differentiator, gaps close with each new release.

3. The scarce resource is the data models reason over,

and the months between visits which are missing from the chart.

4. Ambient AI's $5B+ went to the ten-minute visit.

The 5,000 waking hours a year are still uncaptured. Capturing context has value and most of that context is missed.

5. Epic displaces startups on data it already owns.

It cannot bundle data it does not own and doing so is not zero marginal cost to CIO. Capture the market and distribution before EPIC notices.

Business Model: Severity Captured at the Door.

Severity is earned at the door and lost in documentation.

Downstream tools claim what the chart already proves. Most buried comorbidities never get that proof, so the capture is lost before coding begins. We create the proof at arrival. Proactive care offers doctor feedback that leads to downstream DRG upgrades.

$4,850 / stay

When one MCC moves DRG 293 → 291 (CMS FY2026)

5.8% of admissions

Carry undocumented DRG-shifting comorbidity at mature CDI hospitals (Gabel 2024)

30% denial rate

Initial denial on commercial claims; disputes settle on whether proof was contemporaneous (Kodiak 2024)

Severity value: $0.3M to $1.6M per 50,000 visits, one of three pools.  |  Buyer: CFO signs, CMO approves, ED chair champions.

ROI: Three Pools of Hospital Value.

The hospital keeps most of every modeled dollar.

Per 50,000 ED visits per year. Modeled from published inputs, every assumption labeled. The study sites replace these with measured results.

Value Range by Pool

Total value: $2.4M–$4.4M vs. subscription cost of $0.25M–$0.75M per 50,000 ED visits annually.

Three Value Pools

  • PREVENTABLE RETURNS: 22.4% of ED discharges return within 30 days (Rising 2014); 22% relative reduction in HF hospitalization with between-visit monitoring, 41 trials (De Lathauwer 2025); $12,000 to $15,200 per avoided readmission (HCUP) = $1.0M to $1.3M.
  • THROUGHPUT AND BOARDING: 175 minutes of average boarding (EDBA 2022); $9,700 to $13,300 per boarding hour (Pines 2011); faster disposition when the record is complete at arrival = $1.1M to $1.5M.
  • SEVERITY CAPTURED AT THE DOOR: 157 upgrades that hold per year at a blended $5,000 DRG delta, three of ten assumed rejected = $0.3M to $1.6M.
  • AUREVIA SUBSCRIPTION: Annual departmental contract, 3-year terms; implementation fee $50K to $100K, roughly breakeven = $0.25M to $0.75M.

Scaled examples for four sites:

Yale New Haven Hospital

~190,000 visits → $9.1M to $16.7M

NYP / Weill Cornell

~96,000 visits → $4.6M to $8.4M

Stanford Adult ED

~80,000 visits → $3.8M to $7.0M

UCSF / ZSFG

~75,000 visits est. → $3.6M to $6.6M

GTM & Traction: Clinician-Led, Bottom-Up.

One site's evidence is the next hospital's reason to buy.

$3M

ARR Target

All four systems converted at $750K each in 12–15 months

27

Emergency Depts

Enterprise rollout across all sites in 24–30 months

$20M

Enterprise ARR

Full rollout revenue potential at contract maturity

Distribution Strategy

1

Clinician Champion Identifies the Problem

ED chairs experience the pain point first hand - re-admissions and poor throughput. Aurevia enters through a trusted champion.

2

IRB-Covered Pilot Generates Real Evidence

A signed evaluative pilot produces measured results: throughput, severity capture, burnout scores.

3

First Live Results → First Paid Conversion

The hospital's own data closes the deal. Subscription converts at contract maturity, ~12–15 months post-pilot start. Those sites lead to enterprise roll-outs

4

Published Evidence → Next Hospital's Reason to Buy

Peer-reviewed outcomes from Site 1 become the fast sales cycle at Sites 2–N. The flywheel compounds with each published result.

Pilot Sites & Status

✅ SIGNED

Yale / YNHH

Dr. Melnick

Usability, cognitive load, burnout, adoption, and workflow feedback, comparative feedback - then full deployment.

PHYSICIAN CENTRIC MEASURES

✅ SIGNED

UCSF · SOLVE

Nilpa Shah + Sarkar

Evaluative pilot plus co-development partnership, then deployment.

Equity + Physician & PATIENT-CENTRIC Measures

⏳ FINAL REVIEW · SEPT 28

Stanford

Dr. Bunney · Dr. Kabeer · Dr. Grant (STEPP)

Path: Lakshika → CMO → safety officer → Dr. Dayton → Dr. Dash → STEPP. ROI and study proposal complete; Sept 28 meeting is with program team leads.

Measuring hard financial ROI + re-admissions

🔄 PILOT DISCUSSIONS

Weill Cornell

Dr. Rahul Sharma, Chair of EM

Pilot in active discussion. Champion: Dr. Mitchell Blutt (Consonance Capital).

Measuring hard financial ROI + throughput + re-admissions

"I love that one. I think that's a real win. I've never even heard Epic talk about that sort of stuff. I feel like you could become a real market mover in that space without Epic even realizing it, and then have enough market share that it won't necessarily be easy for Epic to catch up. I resonate with 3 of the MOATs."

— Rohit Sangal, MD, MBA · Yale EM · Chair, ACEP AI Section

Every Category Owns One End. Aurevia Holds Both.



The Direct Ones - and Why We Are Positioned to Win

vs. Epic & EHR Incumbents

Epic owns the record but not the between-visit gap. Clinicians inside Epic still lack structured patient-reported context from the days before arrival. Aurevia fills that gap without replacing Epic - it feeds directly into the workflow Epic already governs.

vs. Aidoc / Viz.ai (AI Decision Support)

These tools act on imaging and clinical signals already in the EHR. They cannot capture what the patient experienced between visits. Aurevia is additive - not competitive - to imaging AI. It surfaces the patient narrative that no imaging result can provide.

vs. Ambient Scribes (Abridge, Ambience, DAX, Suki)

Ambient scribes reduce documentation burden after the encounter begins. Aurevia solves the upstream problem: what does the clinician need to know before the conversation starts? Different moment, different value, no overlap.

vs. Chronic Care Platforms (Apple Health, ChartSpan, Validic)

These platforms aggregate patient data for longitudinal chronic care management - but their output is not surfaced at the ED. Aurevia is purpose-built for the emergency context: urgent, time-compressed, and acting on incomplete information. No existing platform delivers this at the point of care. A different business model and a different buyer.

Team: Built by People Who Have Lived Both Sides of the Problem.

🟠 Full-Time

🔵 Part-Time · 20 Hours/Week

⚫ Advisory

Full-Time

Ammar Ahmed, DO

Founder & CEO

Urgent care, primary care & emergency medicine · Kaiser Permanente residency · FHCSD San Diego

Mehrbod Sharifi

Co-Founder

Senior engineering technical lead · Google DeepMind · Gemini context lead

Rohan Nakra

Product

Full-time product management

Senior Engineer

Engineering

Hire on Close through Mehrbod's network - actively interviewing

🔵 Part-Time · 20 Hours/Week

Nathan Roll

Head of AI

Stanford PhD, NLP · Cambridge · a16z-backed founder

Dr. Eric Basile

Medical Leadership

Chief of Medicine · Columbia

Dr. Lakshika Tennakoon

Chief Scientific Officer

Stanford clinical epidemiologist & research data scientist

Isaac Gutterman

Engineering & Operations

Harvard · MIT

Oscar Schiff

Finance & Strategy

Harvard

Part-Time Product Engineering

Yikuan Sun - Harvard Math & CS · Lunal Graphics

Benjamin Mujkic - IMO medalist · 1× gold · 5× IOI · Harvard

Matthew Chin - Harvard · Machine Intelligence Society


Part-Time Research & Published

Grace Brown - Stanford PhD · Linguistics · Speech perception

Irene Yi - Stanford PhD · Sociolinguistics & clinical language

CLPsych 2026 - Stanford, MIT, Harvard co-authors · 3rd of 17 calibrated presence · 1st summary consistency · 2nd deterioration signatures

⚫ Advisory · Business and Medical

Business: Jeffrey A. Sachs (Sachs Policy Group · President Obama's Health Policy Committee) · Richard T. Miller (Former CFO, NYU Hospital Systems · Former EVP, Northwell) · Tim Peng · Dr. Lisa Masson · Dr. Pamela Resnikoff (VNS Health · Cedars-Sinai, UCLA, FAMIA · Harvard Medical School, Pulm/CC)

Expert ER Advisory Council: 25+ external physicians including Dr. Scott Casey (Kaiser), Dr. Dustin Ballard (Kaiser, RISTRA), Dr. Simon Mahler (Wake Forest), Dr. Frank Peacock (Baylor COM), and physicians from Kaiser, Harvard, Stanford, Cleveland Clinic, MD Anderson, Baylor, UCSF, Mount Sinai, Northwell

The Round: $1.5M SAFE · $10M Post-Money Cap · Pre-Seed.

$1.5M

This round funds the move from validated product to outcome-linked deployment. Everything so far, the platform, the Yale and UCSF agreements, the peer-reviewed result, the council, was built without outside capital, so the $1.5M goes to creating our first commercial deployments to the tune of 3M ARR.

USE OF FUNDS

  • Epic and health-system integration
  • Live clinical and economic validation at pilot sites
  • Product and implementation engineering
  • Patient and care-team operations
  • Security, compliance, and deployment readiness
  • Key hires: implementation lead, clinical operations, a second engineer

WHAT IT REACHES · 10 TO 15 MONTHS

  • Four to six departments live, academic and community, Yale, UCSF, Stanford, Cornell results in hand; the first multi-site outcome evidence
  • The first paid enterprise contracts
  • Patients enrolled through ER discharge at every live site
  • The next raise on evidence rather than on a claim

The defining asset in clinical AI will be generated in workflow: the decisions, outcomes, and patient context that do not exist in any training corpus.