Our promise on trust

Technology should strengthen judgement, never stand in for it.

Care is one of the last places where being unhurried is the whole point. We use technology to give time back and to make decisions clearer — not to make them for you. We publish the rules before we publish the tools.

Trust is not accuracy.

Trust is the confidence that comes from understanding. You trust a colleague who does these six things. We hold every tool we build to the same standard.

Explains itself
Shows its evidence
Admits what it doesn't know
Remembers the context that matters
Shares your values
Helps you avoid a costly mistake

Two principles we keep coming back to

Nobody is paying for technology

People are paying to avoid expensive mistakes. When something changes about a person in your care, you don't need another chat window. You need someone to tell you what changed, why it matters, and what to do next.

Judgement stays human

Technology should strengthen wisdom, relationships, compassion, leadership and community — the things that change slowly and matter most. Every interface decision here is measured against that.

One opinion is never enough

No single view sees the whole picture, so we ask several and show you all of them. The last line is always yours.

01
A colleague's view
What the last person on this route noticed.
02
Expert knowledge
Clinical and care guidance held by the organisation.
03
Personal context
The life story, the preferences, the small things.
04
Community experience
What families and teams elsewhere have found works.
05
Evidence
The moments and notes any suggestion is standing on.
06
Your decision
Always the last word. Nothing happens without you.

Six guardrails

HIPAA before anything else

Every AI tool that touches information about a person requires a business associate agreement and strict protections. No exceptions, no pilots that skip this step.

A human is always in the loop

No AI makes a care decision on its own. Every suggestion is a suggestion, reviewed by the person who knows the human being involved.

Bias and equity are tested, not assumed

Models are validated across race, language and cognitive status before they influence anyone's day.

Consent and dignity are the price of entry

Life stories, voice and any recording require informed consent with an easy, permanent opt-out. A story belongs to the person who lived it.

Pilot before scale

New AI runs in shadow mode for four to eight weeks before it influences a single visit.

Direction comes from the senior team

The AI does not decide what matters. Leadership publishes the joy priorities, and every suggestion cites the one it is following.

What you can see behind any suggestion

Anyone in the organisation can open any suggestion and see exactly this. No part of it is hidden from the people it affects.

Where it appeared: Next Joyful Action
Reasoning shown: In plain words, before you act
Evidence it stands on: A colleague's note, consented and dated
Direction it follows: Every person is known by their story
Set by: Senior leadership team, this season
What it isn't sure about: Stated out loud, not glossed over
Who decides: You. Nothing is marked done without you
Personal health information used: None

Hold us to it.

Every team member can see the trace, question a suggestion, and turn one off. That is the point.