Breaking
$45M Series B closed (2022) Mayo Clinic Platform portfolio company Trained on ~100M patient visits Tens of billions of claims data points AI triage covers 98%+ of member inquiries Revenue tripled YoY since 2020 Blue Cross & Blue Shield of RI on board $45M Series B closed (2022) Mayo Clinic Platform portfolio company Trained on ~100M patient visits Tens of billions of claims data points AI triage covers 98%+ of member inquiries Revenue tripled YoY since 2020 Blue Cross & Blue Shield of RI on board
Healthcare · Artificial Intelligence · Boston

Diagnostic Robotics

Medical-grade AI that reads the data hospitals already have - and predicts who needs care before the emergency arrives.

AI Triage Clinical Prediction Population Health Care Management
Diagnostic Robotics logo
The caduceus, but make it a confidence interval.
Who they are, right now

A patient fills out a short questionnaire. Somewhere, a model is already deciding what happens next.

It is an ordinary Tuesday in an emergency department. Someone describes their symptoms into a form. Before a clinician has read a word, a Diagnostic Robotics model has scored the case, ranked the likely paths, and flagged whether this person is heading toward a hospital bed or a phone call home.

This is the quiet thing Diagnostic Robotics actually does. Not a robot in a white coat - the name is a little theatrical - but software that turns billions of insurance claims and nearly a hundred million prior patient visits into a usable forecast. Health plans and providers use it to find the patients who will get expensive and sick later, and to do something about it now. The company is headquartered in Boston, employs around 66 people, and counts Mayo Clinic among both its partners and its investors.

Most of healthcare is built to react. Diagnostic Robotics is built to predict. - The premise, stated plainly
healthcare aipredictive analyticsvalue-based carerisk stratificationdigital health

The problem they saw

The most expensive disease is the one everyone saw coming and nobody acted on.

Healthcare has a budget problem and a workforce problem, and they feed each other. There are not enough clinicians, the ones who exist are exhausted, and a startling share of spending goes to crises that were visible months in advance - the heart failure that deteriorated, the avoidable ER visit, the mental health decline that went unmanaged.

The data to see these patterns mostly already exists. It sits in claims files and visit records, vast and largely unread. Diagnostic Robotics' founders looked at that pile and saw the obvious, uncomfortable gap: the system collects enough information to predict who is at risk, and then waits for them to show up in crisis anyway. Knowing is not the bottleneck. Acting on what you know is.

Hospitals are drowning in data and starving for foresight. - The gap, in one line
~100MPatient visits used to train models
10s BClaims data points ingested
98%+Member inquiries covered by AI triage
$45MSeries B raised in 2022

Figures reported by the company and press coverage. Treat the round ones as round.


The founders' bet

Three people who believed prediction could be medical-grade, not just impressive in a demo.

Diagnostic Robotics was founded in 2017 by Kira Radinsky, Jonathan Amir and Moshe Shoham. Radinsky had been eBay's director of data science and chief scientist in Israel before most people finish a PhD program's small talk. Shoham is a Technion robotics professor - the academic weight behind the name. Amir rounded out the founding team on the business and product side.

Their wager was specific. Plenty of companies could build a model that predicts illness in a slide deck. Far fewer could build one a clinician would trust at the point of care - validated, explainable, and accurate enough to route a real human toward the right next step. They bet that "medical-grade" was the hard part, and the part worth owning.

KR

Kira Radinsky

Co-founder · CEO / CTO
Ex-eBay chief scientist (Israel)
JA

Jonathan Amir

Co-founder · President / CEO
Product & commercial lead
MS

Moshe Shoham

Co-founder · Engineering
Technion robotics professor
Using the most precise predictive models to drastically improve member clinical care journeys. - Dr. Kira Radinsky, Co-founder

The product

Four pieces, one job: get the right care to the right patient before it costs more to wait.

The platform is less one app than a stack of predictive jobs. A patient answers a simple questionnaire; the system reads symptoms, history and results, predicts the likely course of care, and points them to the right site - ER, clinic, or home. Underneath, models score populations for who will deteriorate, and a virtual agent handles the flood of routine medical questions that would otherwise eat a nurse line alive.

AI Triage Platform

Analyzes a patient questionnaire in real time, predicts the most likely care path, and routes patients to the most appropriate site of care. Built to supplement - not replace - clinician judgment.

Intelligent Care Journeys

Care-management platform with predictive models for avoidable ER visits, congestive heart failure deterioration and serious mental illness, so plans can target interventions where they matter.

Maxine, the Virtual Agent

An AI agent that fields members' medical inquiries at scale - reportedly covering more than 98% of them - and escalates the cases that genuinely need a human.

COVID-19 Triage & Monitoring

A pandemic-era risk assessment and remote-monitoring tool, rolled out statewide to Rhode Island residents and offered to governments worldwide.

The model doesn't overrule the doctor. It hands the doctor a ranked list of who needs them most. - How the triage actually works
Milestones

A short history of seeing it coming

2017Founded by Kira Radinsky, Jonathan Amir and Moshe Shoham to build medical-grade clinical prediction.
2020 · AprPivots fast into COVID-19 triage and monitoring; deployed statewide with the Rhode Island Department of Health.
2020 · JulAnnounces co-development with Mayo Clinic to automate emergency department triage in Rochester, Minnesota.
2020+Partners with Salesforce and Deloitte to offer the COVID platform to health systems and governments globally.
2022 · JulCloses $45M Series B led by StageOne Ventures; becomes a Mayo Clinic Platform portfolio company.
2020-22Revenue reported to have tripled year over year as care-management deployments scaled.

The proof

When Mayo Clinic both deploys your software and invests in it, the demo is over.

Skepticism is the correct posture toward any company that says "AI" and "healthcare" in the same breath. So here is the evidence, not the adjectives. Mayo Clinic agreed to co-develop automated ED triage and later joined the Series B - a customer and an investor are two different kinds of conviction. Blue Cross & Blue Shield of Rhode Island runs care-management programs on the models. During the pandemic, an entire US state pointed its residents at the triage tool.

The $45M Series B in July 2022 was led by StageOne Ventures, with Mayo Clinic, the Technion, and Berkshire Partners co-founder Bradley Bloom among the backers - alongside a roster of Israeli institutions. Total funding sits around $84M. None of this proves the models cure anything. It does prove that serious, cautious buyers were willing to put both patients and capital behind them.

Why "predict" beats "react"

# illustrative scale of the company's claims & coverage figures
~$84M
Total funding ($M)
$45M
Series B ($M)
98%+
Inquiries covered by AI triage
~100M
Patient visits in training data

Bars are scaled for readability across different units, not a like-for-like comparison - the point is the orders of magnitude, not a horse race.

A customer pays you. An investor bets on you. Mayo Clinic decided to do both. - On what real validation looks like

The mission

Make high-quality care scale, precisely because budgets and staffing will not.

The mission is unromantic and, for that reason, credible: relieve strained health budgets and workforces by directing clinically-actionable care to the right patient at the right time. No promise to cure disease. A promise to waste less - less avoidable utilization, fewer crises that a phone call could have headed off, fewer clinician hours spent on cases an algorithm could triage.

There is an honest tension here that the company does not pretend away. Prediction in medicine is only as good as the action it triggers, and only as trustworthy as it is explainable. The bet is that "medical-grade" - validated, auditable, accurate at the point of care - is what separates a useful tool from an interesting one. The competitive field is crowded with healthcare AI and payer-analytics firms; the differentiator Diagnostic Robotics claims is the rigor, not the novelty.

Prevention is a nice word until someone builds the math behind it. - The whole thesis, compressed
Why it matters tomorrow

The aging curve is not waiting for healthcare to hire its way out.

Populations are getting older, chronic conditions are getting more common, and the supply of clinicians is not keeping pace. That math does not resolve itself with effort. It resolves with leverage - with software that lets a fixed number of clinicians reach the patients who need them most, sooner. If value-based care is the direction the industry is heading, prediction is the engine, and Diagnostic Robotics is building engines.

Return to that ordinary Tuesday. The patient fills out the same short questionnaire. But now the form is not a formality - it is the first read of a system that has seen a hundred million visits before this one. The model scores the case, the clinician sees the patient who actually needs them, and the heart-failure crisis that used to arrive by ambulance instead arrives as a phone call three weeks earlier.

That is the change Diagnostic Robotics is trying to make ordinary. Not a robot doctor. A quieter, stranger idea: a healthcare system that calls you first.

The future of care might not look like a breakthrough. It might look like a well-timed phone call. - Where this is all pointed