AI Equipoise

Research & Assessment — AI in Critical Care Workflows

Clinical judgement, weighed against technical capability — not instead of it.

Independent research and assessment for AI entering critical care and mental health workflows. Grounded in clinical practice, built to ask the questions a purely technical or purely commercial review would miss.

Equipoise — the genuine, disciplined uncertainty a clinician holds before acting on incomplete evidence. Applied here to AI, at the exact moment innovation is moving faster than the evidence around it.

What this is

Not a monitoring service, and not another vendor. A research and assessment practice — examining whether AI entering a clinical workflow is built on a genuine foundation, before it reaches a patient.

Who this is for

Clinicians, founders and organisations working at the point where clinical judgement and AI capability meet — and where neither side alone can fully see the risk.

CLINICIANS

Fielding AI pitches, not being asked the right question

Every vendor demo promises efficiency. Almost none ask what would actually make you trust a system with a patient's care.

FOUNDERS

Building something real, and staging it carefully

Teams who've already drawn a line between what's deterministic and what's still probabilistic — and want that reasoning tested by someone outside the building.

INVESTORS

Assessing clinical substance before committing capital

A clinical read on whether a health-AI venture's evidence base and commercial pathway can actually survive contact with real care.

How the assessment works

Four questions, applied plainly, before anyone signs a contract or hands over trust.

01 — Scope

What is actually being decided?

Separating deterministic, rule-based output from genuine probabilistic judgement — and naming exactly where that line sits.

02 — Evidence

What was it built and tested on?

Whose data, whose outcomes, whose definition of "correct" — and whether that evidence holds up outside its own benchmark.

03 — Foundation

Is risk assessment structural, or bolted on?

Whether a genuine biopsychosocial needs assessment sits underneath the system, or whether it was added after the fact.

04 — Continuity

Does it know when something has changed?

Whether the system can tell a bad day from a genuine shift — and knows who to alert when it matters.

Beta · In Development

1Centric AI

An architecture model, developed through this research practice: one person, one continuous AI–patient alliance, built on a foundation that makes it trustworthy rather than just persistent.

FoundationNeeds & risk assessment — owned by the patient, orchestrated by the practitioner.
PathwayDefined by the practitioner. The AI complements it, never overrides it.
AllianceContinuous, contextual, cumulative — no resets.
ContingencyChange is read against the original assessment, and the right specialist is alerted.
Ask About 1Centric AI

Why a clinical view

MB

Most AI risk review is done by people who understand the model but have never sat with a patient, or by clinicians with no way to interrogate the technology. The gap between those two worlds is where real risk hides.

This practice is built on registered mental health nursing experience at the highest clinical specialist grade, over a decade in commercial and international business development, and ongoing study in clinical data modelling and AI in healthcare — applied one question at a time, honestly answered.

Start with one conversation

No pitch deck required on your side. Bring one AI system you're evaluating, or already running, and we'll work through what's actually known about it.