Production AI for regulated finance. Proof the FCA can read.
An engineer joins your team and ships one job in four weeks, inside your codebase. What stays is a record of what the AI was allowed to do, what it did, and who signed it off, for Consumer Duty outcomes, a skilled-person review or your next gap analysis.
The questions land before the evidence exists.
- You have just been asked to feed into a gap analysis, and the evidence is scattered across teams.
- The chief executive does not want to think about it. They want to know it is being looked after.
- Everything you do has to travel upward as a business case, in a planning round that runs once a year.
- Usability testing happens, but not with customers who have the specific needs the rules are about.
- The tooling you bought is a sticking plaster, and everyone on the team knows it.
- AI is doing more of the work, and accountability does not move with it.
The regulator's question is not whether you used AI. It is who allowed what it did, and what you can show.
In. Build. Leave. Prove.
One job, four weeks, in production. Then it keeps proving itself.
In
One job, the rules it has to respect, and the names of the people who sign. Week one.
Build
Our engineer builds it in your codebase, next to your developers, old way and new way side by side.
Leave
Your team runs it. Nothing about the evidence depends on us being there.
Prove
Every decision checked before it happens and recorded, replayable when the question comes.
Three jobs a four-week deployment could take.
- Accessibility fixes on your iOS and Android apps ahead of a gap analysis, with every change checked against the standard before it goes in and a named person signing it off.
- Customer letters and notices produced in the format each customer asked for once, drafted by AI and released only by a named handler.
- Triage of inbound complaints and vulnerable-customer flags, with the draft response and the evidence for the board report assembled, and a named handler releasing every one.
- Whichever we start with, the record is the same: what was allowed, what happened, who signed.
Who approved this version, on what evidence, under whose authority, and can you show it was not altered?
What the AI is allowed to do is written down first.
Every decision is checked against those rules before it happens.
A named person signs it off. The check produces the evidence; a person judges.
You can replay the whole history any day and get the same answer.
If anyone changes it later, it shows. Patent pending, UK application GB2620101.2.
That is what a deployment leaves running for your job. In the accessibility product today, a person on your team accepts every finding before it reaches your record.
Who this is for.
- Digital and customer-experience leaders who own the app and the logged-in journey.
- Compliance and risk leaders who hold a personal regulatory responsibility for what the firm does.
- Teams who have been asked for evidence and found it scattered across inboxes and spreadsheets.
Not for
A compliance report with nothing running behind it. This starts with a job in production.
Two people, on every call and in your standup.
Simon Milner, Founding Architect
He designed the record: what the AI is allowed to do, checked before it acts, and replayable afterwards. Twenty-five years in Silicon Valley before that.
Jason Crispin, Founder
He owns the customer side of every deployment: what the job is, what it is worth, and that it lands. He is on the first call and every one after.
Patent pending, UK application GB2620101.2. Meet the team
Four weeks, then it keeps proving itself.
Week 1
The baseline
What the job is, what allowed means for it, and who signs. Written down before anything runs.
Weeks 2 to 4
The build
Our engineer works in your codebase next to your developers. The old way and the new way run side by side.
Week 4 on
The proof
Every decision checked and recorded. Replay it any day. We maintain it, or you run it without us.
What you keep
- The job, live, in your product.
- The code, assigned to you in writing.
- The record of every decision, replayable any day.
- The check, still running on every release.
What people ask.
Does Consumer Duty require an accessibility audit?
Does Consumer Duty require an accessibility audit?
No. Consumer Duty is about outcomes, including outcomes for customers in vulnerable circumstances, and it does not name WCAG or any digital standard. What it does mean is that when someone asks how you know those outcomes are good, you need evidence rather than an operational metric used as a proxy.
Who signs off what the AI does?
Who signs off what the AI does?
People you name, at the points you choose, before the action takes effect. Accountability does not move to the software: the check produces the evidence and a named person judges.
Our logs are already good. Why is this different?
Our logs are already good. Why is this different?
A log records what happened. It cannot show that each action was allowed at the time, on what evidence, or that the record has not been changed since. That is the difference between a log and a record you can replay.
Can our skilled person or auditor see it?
Can our skilled person or auditor see it?
You can replay the history and show the same answer any day: what was proposed, what was allowed, who allowed it and when. There is no separate auditor login yet, so today you show it to them.
How do you start without a long procurement cycle?
How do you start without a long procurement cycle?
One job, four weeks, one engineer inside your team. It is small enough to scope on a call and to carry as a business case into a planning round.
What does it cost?
What does it cost?
Scoped on the call, because it depends on the job.
What have you been asked to prove?
Thirty minutes with Simon and Jason. Bring the gap analysis, the board question or the job you have been asked to hand to AI, and you leave knowing what it would take and what it costs.