Six ways in. One record underneath.
Pick the job. We put an engineer inside your team, ship it in four weeks, and leave a record that proves every decision the AI made was allowed.
Six places to start.
Four are scoped on the call. Accessibility is live today, and document sign-off is in early use.
AI-Native Delivery
Your roadmap, shipped the new way. Our engineer runs your delivery with agents from ticket to pull request, inside your codebase, and leaves the setup behind.
Scoped on the call
See AI-native deliveryCustom AI Agents
One agent in one process, with what it is allowed to do written down first, a named person on every consequential action, and a record you can replay.
Scoped on the call
See custom AI agentsAI Harness Engineering
The control layer around the agents you already run: evaluations, pre-action checks, approval gates and replay, wired into your CI.
Scoped on the call
See harness engineeringSovereign AI Deployments
Models, data and the record on hardware you own or a UK cloud you control, for work that cannot go to a US-hosted model.
Scoped on the call
See sovereign deploymentsAccessibility
Guided checks for every page and screen against WCAG and EN 301 549, a ticket for every failure, and a statement built from the results. Websites free.
See accessibilityDocument Sign-off
A named reviewer and a named approver on every document, with a history an auditor can follow. In early use.
See document sign-off
In. Build. Leave. Prove.
One job, four weeks, in production. Then it keeps proving itself.
In
You bring one job. We agree what the AI is allowed to do in it, and who signs.
Build
Our engineer works in your codebase next to your developers, old way and new way side by side.
Leave
The job runs in production and your team runs it. The check and the record run on AUDITSU, on a monthly subscription.
Prove
Every decision is checked before it happens and recorded, and you can replay the history any day.
Bring one job.
Thirty minutes with Simon and Jason. You leave knowing what it would take, what you would keep, and what it costs. If the honest answer is not to use AI for it, we will say so.