Ethos
How your organisation decides what AI is for.
Every organisation is deploying AI. Almost none have decided what it's for. Ethos guides your leadership team through that decision, producing an organisational AI constitution, a deployable system prompt for every AI tool you run, and something rarer than either: a leadership team that actually agrees.
The decisions nobody has made.
Should AI challenge your people's thinking or execute their instructions? Produce more output, faster, or fewer outputs with greater impact? Answer every question, or connect people to the colleague who could sharpen their thinking? These are leadership decisions, not technical settings. In most organisations, nobody has made them.
How it works.
Each leader responds blind. Strong positions are gently interrogated, because a position without a reason can't govern anything. Then the room sees where it agrees, and where it doesn't. The disagreement is the point: Ethos measures alignment before and after the conversation, argues the minority view so nobody has to be brave alone, and forces a decision on what wins when principles collide.
What you leave with.
Your AI Constitution, with every principle carrying its reasoning. A deployable system prompt for Claude, Copilot, or any enterprise AI tool. A divergence report showing how your team's alignment moved. And a one-page governance summary your board can table as evidence of real AI oversight.
Your AI tools already have values. The only question is whether they're yours.
Two ways to run it
Facilitated workshop, or run it async.
Live workshop. A facilitator sets the room up, participants answer blind, then the team meets to reveal alignment, argue the hard cases, and ratify together.
Async study. No facilitator available? Add your leaders, send magic links, and each one walks themselves through the five stages on their own time — voice input as they think, private brainstorm pad per stage. You synthesise when enough voices are in.
What to expect
A guided process, not a workshop you have to run.
Ethos is the instrument. It carries the structure, asks the questions, and produces the artefacts. Your team brings the judgement.
- 01
Set the context
Your facilitator adds the organisation's values, strategy, and any existing AI policy. Ethos grounds every question in your reality.
Facilitator · 10 min
- 02
Answer blind
Each leader works through the stages alone. Ethos asks for reasoning, not just positions, and gently interrogates weak arguments.
Each participant · 30 min
- 03
See the room
The facilitator opens the alignment reveal. The team sees where it agrees, where it doesn't, and where the disagreement actually is.
Team · 60–90 min
- 04
Leave with artefacts
Ethos synthesises your AI Constitution, a deployable system prompt, a divergence report, and a one-page governance summary.
Delivered · same day
For participants
Around 30 minutes, on your own, at your own pace. You answer blind. Ethos may push back on one or two answers to make sure your position has a reason behind it. Your inputs stay private until the room comes together.
For facilitators
Set the organisational context once. Invite the leadership team by magic link. Run live or asynchronously. When the room is ready, present the alignment reveal and let Ethos draft the constitution.
Participants
Join a session
Your facilitator will share a six-character code, or a personal link sent directly to you.
Facilitators
Run a session
Sign in to create an organisation, invite your leadership team, and open a session in live or asynchronous mode.
Pilot programme
Run Ethos with your leadership team.
We're partnering with a small number of organisations to run foundational sessions during this pilot. You bring the leadership team and the context. We help you facilitate, and both sides learn from what the room surfaces.
Confirm the room
Six to twelve leaders. The people whose judgement should shape how AI shows up in the organisation — typically the exec, plus one or two dissenting voices.
You · 1 conversation
Prepare the context
Values statement, strategy summary, and any existing AI guidance. If you don't have one yet, that's a finding, not a blocker.
Facilitator · 30 min
Run the sequence
Participants complete the pre-work asynchronously over 3–5 days. The reveal, synthesis, and ratification happen in a single 90-minute session — live or over video.
Team · 1 week end-to-end
What we ask of pilots
- · A named facilitator inside the organisation.
- · A 30-minute debrief with us after the session.
- · Permission to use the anonymised patterns (never text, never names) to improve the instrument.
Privacy
Individual answers are never shown to other participants until the facilitator opens the reveal. Nothing you write is used to train third-party models. Your organisation's context and the resulting constitution belong to you.