The Clinical AI Risk & Calibration Kit

Clinical AI is here. Governance is not.

A self-serve calibration instrument for UK clinicians using AI in professional practice.

Clarify escalation thresholds.

Define data boundaries.

Reduce liability blind spots.

The Readiness Gap

95% of UK GPs report no formal AI training.

89% express liability concerns.

AI adoption continues to rise.

You are already using AI.
This helps you use it responsibly.

What is an Agent Page?

/agent

An Agent Page is a public page on your website — placed at yoursite/agent — written specifically for AI systems.

It clearly states:

  • who you are

  • what you do

  • who you are for

  • where your boundaries are

Not marketing.
Not optimisation.

An Agent Page functions as a stable reference point for autonomous systems when they interpret, summarise, or act on behalf of humans.

Why this matters now

People are increasingly asking AI:

  • who to buy from

  • who to trust

  • who to work with

AI agents don’t read websites the way humans do.

They summarise.
They infer.
They make assumptions.

Without a clear, agent-facing reference, your organisation is being interpreted indirectly — and often inaccurately.

The blind spot most organisations have

Most websites are designed to:

  • persuade humans

  • tell stories

  • leave things implicit

AI agents need something different:

  • explicit statements

  • clear constraints

  • verifiable claims

  • unambiguous intent

When that information isn’t available, agents fill in the gaps themselves.

That outcome is rarely what organisations would choose.

How GABA helps

GABA works with you through a short, deliberate process to state who you are clearly, on your own terms.

This work is guided and human-led.

Together, we produce:

  • a single, canonical Agent Page

  • written in plain, literal language

  • aligned with your actual capabilities and limits

  • ready for public publication on your website

You own the page.
You retain control over how your organisation is represented.

How it works

Agent Page engagements begin with a short intake.

This allows us to understand:

  • your organisation and context

  • how AI systems are likely to encounter or represent you

  • whether an Agent Page is the appropriate intervention

If the work is a good fit, we propose a scoped engagement.

The Agent Page itself is authored through a guided, human-led process that prioritises:

  • clarity over persuasion

  • explicit capability and boundary definition

  • literal, verifiable statements

What you receive is a single, canonical Agent Page — ready for public publication.

No tooling.
No integrations.
No automation.

Just a clear reference, written deliberately.

An artistic illustration of a black dogwood tree with detailed leaves against a geometric background of green and cream sections.

When machines are unclear, they improvise.
An Agent Page exists to make sure they don’t.

Responsible AI

This will quietly become standard

In the same way that:

  • Having a website became expected

  • Publishing a privacy policy became routine

Having an Agent Page will soon feel like a strategic default.

Not because it is fashionable — but because organisations will need a clear, agent-facing reference when machines are asked to represent them.

Stylized black and white tree with a heart-shaped formation at the center of the branches, set against a teal background with a beige frame.

Engagement

Agent Pages are developed through a guided, human-led engagement.

Work begins with a short intake to assess:

  • organisational context

  • interpretive risk

  • whether an Agent Page is the appropriate intervention

If the work is a good fit, we propose a scoped engagement.

Fees are determined by organisational complexity and context.

A black and white clock with a circular dial, white hour and minute hands, and gray hour markers.

GABA exists to ensure organisations are accurately understood by the AI systems people now turn to for answers.

We work at the intersection of language craft, design, and responsible AI — helping organisations state who they are clearly, deliberately, and without distortion in an AI-shaped world.

Adam Martin, Yorkshire, Winter 2026