Experimental / v0.9 • 2026
An innovation initiative by AST

Ambient AI that writes
the clinical note for you.

Medexa listens in the background and turns a visit into compliant documentation and billing codes — in real time, at the point of care. No dictation, no catch-up hours.

Meet Medexa → How it works
0s
Post-visit documentation
8-min
Rule automated
Mod-59
Compliance built in
An innovation initiative by AST

A copilot that speaks
clinical & billing.

Ask our AI Medical Copilot anything — from "how do I add a modifier?" to "what's the policy for a Medicare Advantage claim?" — and get a grounded, cited answer in seconds.

Try the Copilot → Read the docs
<2s
Median answer time
100%
Cited to policy
24/7
In production now
An innovation initiative by AST

Compliance, wired in
from day one.

HIPAA, ONC, and payer rules aren't audits at the end — they're a design constraint from the first line of product code. 17 years of healthcare IT tell us the difference.

See our approach → Talk to us
17+
Years in healthcare IT
98%
Client retention
380+
Engagements delivered

Bringing AI-native thinking
to the workflows clinicians use every day.

AST Labs compounds decades of healthcare IT delivery into focused products that ship — not demos.

2
Products in-flight
100%
Healthcare focus
Day 1
HIPAA alignment
US
Clinician-validated
Our Products

What we're building.

Two focused products today. Each one rebuilds a specific, costly clinical workflow from the ground up.

Ambient Clinical Documentation In development

Medexa medexa.ai

Ambient note generation, 8-minute rule automation, and 59-modifier compliance — built for therapists, not generic EMR users.

Rehab Therapy CPT / ICD-10 8-min rule Mod-59
Latency < 800ms · stream-to-note
Stage Design-partner pilots
Join the pilot →
Conversational Workflow Assistant Live

AI Medical Copilot copilot

A conversational assistant trained on clinical workflows and EMR operations. Answers are grounded in policy manuals, payer guides, and your org's own documentation.

Grounded RAG EMR-aware Cited answers Workflow tasks
Response 2.3s · first token
Stage In production · 4 orgs
Open the copilot →
The Lab

Work in progress.

A public-ish view into what's on our bench right now — active research threads, beta experiments, and early pilots. Some will graduate to products. Some won't.

Research LAB-R04

Low-latency ambient transcription for noisy clinical rooms

Evaluating speaker-separation models at 16kHz under real PT-clinic noise floors. Target: < 400ms first-token, < 2% WER on domain terms.

Q1 2026 · ongoing
SRJKAT
Research LAB-R07

Self-correcting CPT coder with reviewer feedback loops

Training a lightweight reranker on billing-supervisor corrections so the model learns each clinic's local rule quirks without retraining from scratch.

Paper · in draft
MNDP
Research LAB-R12
ICD-10 LCD NCCI payer

Unified clinical knowledge graph for payer-aware reasoning

Merging ICD-10, CPT, LCD/NCCI and per-payer policy into a single queryable graph so downstream models can justify every code with a citation chain.

Q2 · internal
SRDPEW
Beta LAB-B02
/voice-note → 97110 · 97530 · mod-59

Voice-note → coded claim, one tap

A mobile-first flow for PTs on home visits: record 30 seconds, get a compliant draft claim before you reach the car. Closed beta with 9 therapists.

Closed beta · 9 users
JKMN
Beta LAB-B05
denial: 59 → missing docs auto-appeal ✓ SLA: 48h

Denial-to-appeal auto-drafter

Reads a denial EOB, pulls the original note, drafts the appeal letter with policy citations. Human-in-the-loop for now; fully-drafted for 71% of cases.

Internal beta · 2 clinics
DPATRK
Beta LAB-B08
EMR Copilot FHIR R4 HL7 v2

Read-only FHIR bridge for Copilot grounding

Connecting the Copilot to patient context via a zero-trust, read-only FHIR proxy. Answers can cite the actual chart — not a summary of it.

Private beta · 1 partner
EWSR
Pilot LAB-P01

Medexa at a 14-clinic PT network

Six-month pilot measuring: time-to-close-note, Mod-59 compliance rate, and clinician NPS. Week-over-week: note-close time ↓ 42%.

Month 4 of 6
JKSR
Pilot LAB-P02
AVG response · 2.3s Resolved · 84% Escalated · 11% 84%

Copilot deployed to a 400-seat billing ops team

Replacing a wiki + internal help desk. Resolving 84% of "how do I…" questions on first answer, with audit-logged citations.

In production · expanding
MNDPEW
Why AST Labs

Built by clinicians and engineers, inside an 11-year healthcare IT shop.

Talk to us →
Our Approach

Serious healthcare AI, built from the inside out.

AST Labs is the healthcare AI unit of All Star Technology — a company that has spent 17 years inside US healthcare IT, building the unglamorous, high-stakes systems that clinicians and billing teams depend on every day.

Labs exists to compound that experience into AI-native products: focused, clinician-validated tools that fit into the workflows clinicians actually use, with compliance and accuracy treated as design constraints, not marketing claims.