Healthcare

AI-native healthcare,
built by people who have shipped it.

Health companions, clinic operations, and digital-health R&D, all on one agentic stack. This is work we have run ourselves, the regulated and operationally heavy kind, with a clinician kept in the loop wherever a decision reaches a patient.

Talk to us See the work ›

Why us in healthcare

The hard part of healthcare is the part we have already done.

The regulated, data-heavy work with a clinician in the loop is where most teams get stuck, and it is the part we are most at home in. We have led digital-health R&D end to end, designed an in-app assistant for one of Israel's largest HMOs, and built the intelligence layer behind a private clinic.

We have run it

We have run the R&D

We led R&D end to end for a US digital-health platform and brought its monthly burn down by more than half, with the product still shipping the whole way through.

Regulated by default

At home with the scary parts

Clinical data, member PII, EMR and pharmacy integrations, compliance audits, device-grade process. This is usually the part that scares teams off, and it happens to be the part we are used to working in.

Clinician in the loop

Agents carry, people decide

The agents carry the volume, the intake, the follow-ups, the data, the first drafts, while the clinician signs off on anything that actually reaches a patient, so the judgment always stays with a person.

What we build

Four places we put the stack to work in health.

Each of these runs on the same five-layer agentic engine, wired into your systems of record, with people kept on the decisions that matter.

Companions

AI health companions

Conversational agents that coach, log, remind, and explain, grounded in each person's own clinical history so the guidance actually fits them.

Clinic ops

Bionic clinic operations

Intake, patient status, records, scheduling, and follow-ups, all run by coordinated agents, so a single physician's judgment can reach far more patients than their hours allow.

0 to 1

VP Product as a service

Senior product leadership when you need it. We stand up a new health product or line in a few weeks and hand it back to your team, without you taking on another full-time hire.

Transformation

Health-org AI transition

We take an established health organization AI-native without breaking what already works, reading the operation board-grade and rebuilding it on the stack while the work keeps running.

Selected work

Three builds, one operating model.

A consumer health coach, an HMO member assistant, and a concierge clinic. Three very different builds that run on the same agentic stack, with the same habit of keeping a clinician in the loop.

Digital health · United StatesR&D run end to end

A personal AI coach for Americans living with diabetes

A US platform helps people manage type-2 diabetes through continuous glucose monitoring and coaching. We took over its R&D end to end, rebuilt the product around a conversational health companion, and stood up the clinical-data engine behind it. Over that stretch the monthly burn came down by more than half, and the product kept shipping the whole time.

The agent logs meals, explains a glucose spike, walks people through onboarding, and sends the reminders the old product used to drop, and it works from each person's actual glucose history while it does.

>50% burn cutmonthly R&D burn, while shipping

960K+ readingscontinuous glucose data ingested via Terra, Apple Health, and FreeStyle Libre

17-tool agentone conversation, built by a four-person pod in eight months

Health fund · IsraelProduct design + build

A personal health assistant for one of Israel's largest HMOs

When a member leaves the doctor they walk out with a stack of follow-ups, renew a prescription, book the right blood test, see a specialist, often in a particular order, and a lot of them get lost somewhere in the middle and never finish. We designed and built an in-app personal assistant that picks the member up after the visit and walks them through the whole thing.

Behind one plain-language conversation, a set of agents does the renewals and the bookings, wired into the fund's records, scheduling, and pharmacy systems.

Post-visitit comes to the member right after the visit, while it is still fresh

One assistantprescriptions, tests, and appointments in plain language

Agents behind itbooking and renewing on the member's behalf, on the existing app login

Private clinic · IsraelBionic clinic design

A bionic clinic for a concierge physician

A boutique clinic where patients pay for one physician's judgment, reachable morning to night, much of it over chat. The doctor had become the bottleneck of his own practice, and his records system was slow and desktop-bound. We designed a phone-first intelligence layer over his existing records.

Type a patient's name and get a one-screen status, latest labs, vitals, medications, recent messages, and an AI summary of the raw documents, then reply in seconds. The aim is a bionic doctor who treats more patients without losing the personal touch they pay for.

One screenlabs, vitals, meds, and messages in one place, where it used to take a slow login and a hunt

Phone-firstbuilt for a physician who works from his phone all day

Human-in-the-loopthe physician edits and approves every message before it goes out

Health and regulated work the founders have built at, with, and for

Sweetch Maccabi Hello Heart Walmart Prospera

Two decades building in regulated, operations-heavy industries.

Put the stack to work in your health business.

Tell us where you are starting, a product, a clinic, or a whole operation. We bring the stack and the team.

hello@3peaks.dev