AugmentedAI · ecosystem online

The AI ecosystem
built for healthcare.

Thirteen CareAI products, thirteen specialised agents, one inference engine — a complete intelligence layer for healthcare. We don't replace the way care is delivered. We augment every step of it.

Talk to us→See how it works
13
CareAI products
13
specialised agents
1
engine · since 1962
Built on healthcare standards
FHIR R4HL7 v2SNOMED CTICD-11LOINCDICOMHIPAAGDPRISO 27001SOC 2 Type IINHS DSPTOpenEHRFHIR R4HL7 v2SNOMED CTICD-11LOINCDICOMHIPAAGDPRISO 27001SOC 2 Type IINHS DSPTOpenEHR
00 / Thesis · since 1962

Augmenting human intellect.

Douglas Engelbart · 1962

Sixty years on, the right place for AI is still alongside human judgment — never in its place.

Replace

Cut humans out of the loop. Hide the model. Outsource judgment.

Augment

Put expertise at the centre. Surface the why. Hand the decision back to the practitioner.

Brynjolfsson 2022 · The Turing Trap
01 / How it works

Products on top.
Specialists between.
One engine underneath.

Every interaction in a CareAI product — a clinician's note, a pharmacist's check, a patient's question — is routed to the right specialised agent. Every agent thinks on the same inference engine. Three layers, one ecosystem, one experience.

CareAI products
CareAI.MD
in
CareAI.Meds
in
CareAI.PBM
in
CareAI.Pharmacy
in
CareAI.Pulse
in
CareAI.HealthFund
in
PrognosisAI engine
prognosis.core
routing · live
Current dispatch
CareAI.MD62ms
AugmentedAI agents
out
Clinical Reviewencounter assist
out
Drug Interactioninteraction check
out
PA Reviewprior-auth decision
out
Formulary Agentformulary check
out
Adherence Coachpatient nudge
out
FWA Detectionclaim screen
01.5 / A moment in care

What changes when you augment?
Watch one Tuesday morning.

One patient. Ten minutes. Three small moments where the ecosystem quietly fires — and one moment where the human stays in charge of the call.

CareAI.MD
MR · 14872
08:14
local · clinic
08:14
human
08:17
Clinical Review
08:22
Drug Interaction
08:24
human
Current dispatch
Practitioner — in the loopsigned
  1. 01
    08:14

    A patient walks into the clinic with a routine medication refill.

    CareAI.MD opens the chart. No agent has fired yet.
  2. 02
    08:17

    A subtle pattern in their labs would have taken twelve minutes to spot.

    Clinical Review Surfaces it in 4 seconds, with three citations to the patient’s own history.
  3. 03
    08:22

    The new prescription would interact with something they’ve been on for years.

    Drug Interaction Intervenes before the script is signed.
  4. 04
    08:24

    The clinician decides.

    No agent. The human is in the loop. Every recommendation reviewed, accepted, modified or overridden by the practitioner.
02 / The products

Thirteen products.
One ecosystem.

CareAI is the surface of AugmentedAI — purpose-built apps that meet clinicians, pharmacists, paramedics and patients exactly where they already work.

CareAI.MD · Medical Practice
01

Every patient prioritised against their full clinical history.

P1Chest pain · 58MSTEMI rule-out
P2Dyspnea · 71FCHF flare
P3Lab trend · 44MAKI watch
P4Med refill · 62Froutine
CareAI.Pulse · Remote Monitoring
02

Continuous signals, watched in real time.

HR 78 bpmSpO₂ 97% sinus
CareAI.MentalHealth
03

Behavioural health, clinically supervised.

Headache since morning, 6/10.
CareAI.Clinic · Clinical Practice
04

Reasoning grounded in the patient graph.

patientCKDT2DMlisinoprileGFR↓UACR↑nephro
CareAI.Meds · Medication Management
05

Interactions and dosing across every list.

Lisinopril×Spironolactone
K+ ↑
Warfarin×Fluconazole
INR ↑↑
Metformin×Iodine contrast
hold 48h
CareAI.Emergency
06

EMS dispatch, forecast hourly.

92%
occupancy
+ seven more in the line-up
CareAI.PBM
Pharmacy Benefits
CareAI.Pharmacy
Pharmacy Operations
CareAI.Baby
Maternity & Paediatrics
CareAI.MedAdvice
Medical Advice
CareAI.HealthFund
Health Fund
CareAI.Obesity
Obesity Programme
CareAI.Roadside
Roadside Assistance
03 / The agents

Thirteen specialists.
One common brain.

Each AugmentedAI agent is purpose-trained for a specific clinical or operational job. They share one inference engine, one memory model and one set of guardrails — so the answer is always consistent, no matter which product asked the question.

01

Chat Assistant

The conversational front door for every CareAI product.

02

Claims Adjudicator

Decides claims against policy in real time, with reasons.

03

Clinical Review

Reviews encounters against guidelines and patient history.

04

Drug Interaction

Catches risky combinations before a script is filled.

05

PA Review

Resolves prior authorisations in seconds, not days.

06

Formulary Agent

Picks the on-formulary alternative, automatically.

07

FWA Detection

Flags fraud, waste and abuse the moment it appears.

08

Prescriber Intelligence

Patterns, outliers and benchmarks across every prescriber.

09

Population Health

Cohorts at risk, surfaced before they become claims.

10

Adherence Coach

Nudges patients toward the next dose, the next visit.

11

Therapeutic Substitution

Safer or cheaper alternatives, clinically equivalent.

12

Document Analysis

Reads referrals, labs and notes — turns them into structure.

13

Member Engagement

Personalised outreach across every channel a member uses.

04 / The engine

The PrognosisAI engine.
Quietly powering it all.

All thirteen agents and all thirteen products run on a single inference engine, purpose-built for healthcare. It's not the thing customers see. It's the thing that makes everything else possible.

01

Multi-model LLM

A panel of frontier and domain-tuned models. The right model picks up the right query — and adjudicates consensus before any answer leaves the platform.

LLM panel4 active
Consensus ruleκ ≥ 0.85
Dissent escalationhuman-in-loop
02

Graph traversal

Patient-level knowledge graphs built from FHIR, SNOMED CT and LOINC. Every reasoning step is grounded in ontology — auditable, traceable, citeable.

OntologiesSNOMED · LOINC · ICD-11
Avg nodes / patient1,427
Edge provenance99.4%
03

Context memory

Agents remember the patient, the encounter and the pathway — within strict consent boundaries. Memory is scoped per agent, per consent, every read audit-signed.

Scopeagent · patient · consent
Retentionpolicy-bound
Every readaudit-signed
04

PII guardrails

Personal data never leaves your boundary. Every model call respects regional and tenant scope — and PHI is filtered at the routing layer, not the UI.

PHI boundarytenant · region
Egress filterrouter-side
Redactiondefault-on
05

Token optimisation

Smart routing, semantic caching and adaptive context windows cut inference cost by an order of magnitude — making the economics work at population scale.

Cache hit rate64.2%
Avg cost / call↓ 11.4×
Adaptive windowauto
05 / Privacy & safety

Built for regulated
environments.

AugmentedAI was designed for healthcare from day one. Privacy, auditability and clinical safety aren't features bolted on — they're enforced by the engine itself, on every call.

HIPAAGDPRISO 27001SOC 2 Type IINHS DSPTDPIA-ready

In the lineage of Engelbart (1962) · Licklider (1960) · Shneiderman (2020) · Topol (2019).

PHI stays inside the boundary

Data never leaves your region or tenant. Every model call, every memory read, every tool invocation respects explicit data boundaries.

Consent-scoped memory

Agents see only what the patient, provider and policy permit. Scopes are enforced at routing time, not at the UI.

Auditable by default

Every dispatch is signed, every reasoning step is traceable, every output is citeable against primary evidence.

Clinical oversight, always

High-stakes decisions carry a human-in-loop requirement. Autonomy is a spectrum the provider controls.

Partner with AugmentedAI

Ready to augment
the way care is delivered?

We work with healthcare providers, payers and investors to bring the AugmentedAI ecosystem to market — on your infrastructure, under your compliance, at population scale.

Talk to usSee how it works