UNCAPT

Expert reasoning,
at scale.

Over a billion people can’t access the experts they need. The experts themselves can’t scale. We built the infrastructure to capture expert reasoning, govern it, scale it, and compound it.

Get in touchHow it works →
Est. 2020 · Sydney
ISO 27001:2022 CertifiedISO 27001:2022
Certified
Brain & Mind CentreMRFFUNSWUniversity of NewcastleiLAKeepableRoyal Flying Doctor ServiceBrain & Mind CentreMRFFUNSWUniversity of NewcastleiLAKeepableRoyal Flying Doctor Service

Clinical validation

0%
triage accuracy
vs expert panel · Brain & Mind Centre, USYD
0
hallucinations
clinical evaluation · Brain & Mind Centre, USYD
0%
BGP perception accuracy
vs GPT-4o's 56% · UNCAPT · under review
0%
reduction in planning time
2 hrs → 15 mins · University of Newcastle evaluation
0
products
one platform · health & research

The problem we’re solving

A billion people.
A handful of experts.

The world’s best judgment sits inside a handful of people. Generic AI trains on what was documented — not on how they reason. It can’t be corrected by the people who know better, and it has no way to keep them in the loop when it’s wrong. UNCAPT is the infrastructure layer between the expert and everyone who needs them. The experts stay at the frontier. We handle everything else.

The category didn’t exist.
It does now.

Five products.
One platform.

Each co-built with a domain expert.

Brain & Mind Centre
MIAMental Health
94%
accuracy vs expert panel

Clinician-quality triage and care planning at scale.

Co-built with the University of Sydney Brain & Mind Centre · miahealth.com.au

UNSWBrain & Mind Centre
InsightResearch Acceleration
∞
living system

Every researcher working at the frontier of their field.

A living knowledge graph built from research literature

Royal Flying Doctor Service
MATILDAEmergency Medicine
86%
BGP perception accuracy

Expert clinical eyes on every case.

vs GPT-4o's 56% · UNCAPT · under review

iLAUniversity of Newcastle
CARAAged Care
87%
time saved

A support plan built around a person's actual life.

From 2 hours to 15 minutes · University of Newcastle

UNSWMRFF
NADIAIntellectual Disability
∅
never starts fresh

Support from a system that actually knows them.

Longitudinal support for people with intellectual disability · builds understanding over time

Why domain leaders work with us

Their expertise scales without leaving their hands. They get to keep working at the frontier.

They rarely consult on the same case twice — the system handles scale while they stay at the frontier.

Our co-build model →

Proprietary IP  ·  Six components

Six components.
One loop that compounds.

Six components built in-house. Every expert correction makes it more capable. Every case it sees deepens what it knows.

01
ELI Platform

Turns uncertainty into the training signal.

Deep-dive →
02
Living Knowledge Banks

Research as a versioned, queryable knowledge graph.

Deep-dive →
03
Stage Verifiers

Decision thresholds learned from expert corrections.

Deep-dive →
04
Governance Layer

Config, verifiers, and regression tests — each updateable independently. Full audit trail on every output.

Deep-dive →
05
Belief Guided Perception

Intake and discovery that sharpens with every cycle of reasoning.

Deep-dive →
06
Domain Orchestrator

Runs the OODA pipeline. Manages state between stages and handles the verifier signal at each gate.

Deep-dive →

ISO 27001 certified  ·  Azure AU  ·  Full audit trail  ·  Data residency controls

Technical deep-dive →

If expertise is the bottleneck,
we should talk.

We work with research groups, hospital networks, and specialist institutions. Every engagement starts with understanding the domain — what knowledge exists, who holds it, what would change if it could scale.