Start small. Prove it. Then grow.
You don't need to know what to build. We find where AI agents would help, try them on real work, and give them more to do as they earn it.
AI-generated image
Blueprint. Prove. Scale. Embed.
| Step | What you get | What must be true before the next step |
|---|---|---|
| Blueprint. | A short, paid first step where we map where agents would help and agree what success would look like. | You've agreed which work comes first, what success looks like, and how it will be measured. |
| Prove. | We build for one piece of real work and show you it working. | The agreed measure is tracked on real work, and you've seen what it shows. |
| Scale. | More work, more teams, more of the business connected. | The first workflow runs inside your operations under written rules, with its level of oversight agreed. |
| Embed. | The aim is that it becomes part of how your business runs, and we keep running it with you. | — |
Four things we do, under one roof.
An AI agent is AI that does a piece of work for you. Getting one to work inside a real business takes more than the AI.
Understand the work.
We map how your business really runs: who does what, where information lives and where time goes.
Design it for people.
We design the screens and each agent's job around how your team works, so the platform fits the way they already work, at a desk or on a phone.
Build it into your business.
One AI platform with each agent given one job and clear rules, connected to the systems you use and put to work in your daily operations.
Run it with you.
We stay on after launch to keep it running and improve it as your business changes.
Freedom is earned.
Think of a new hire on probation. In week one you sign off everything; later you check the things that matter most.
Autonomy, earned.
Our method
What has to be true before oversight steps down.
We publish our thresholds because a buyer should be able to test them. These are our starting values. The method is new, and we'll tune them as real run data builds up.
| Level 1 to level 2 | Level 2 to level 3 |
|---|---|
| At least 4 weeks and 200 reviewed runs at level 1. | At least 8 weeks at level 2. |
| Fewer than 5% of outputs changed or rejected by a person. | Exceptions on fewer than 2% of runs. |
| No serious errors. | Sampled checks at least 98% correct. |
| Written exception rules, tested against past runs. | Value on track: at least 100% of what was expected at that point in the project. |
| At least 70% of planned use actually happening. | Drift alerts and a way to roll back, both live and tested. |
You co-sign every step down.
Each move down a level changes the risk your business carries, so it isn't ours to decide alone. Our delivery lead and your process owner both sign it. For our first workflows at level 3, our CEO signs as well.
Oversight steps back up automatically.
A workflow moves up one level straight away, and the move is reviewed afterwards, in any of three cases:
- there is a serious error;
- corrections or exceptions run above their threshold for two weeks;
- the value delivered falls far enough behind the agreed plan to trigger a recovery review.
You're told within one business day. It steps down again only by meeting the thresholds and being co-signed, as before.
Some decisions never leave a person.
No level of evidence moves these to an agent:
- commitments that change scope, price or risk;
- financial changes above agreed limits;
- legal or regulatory exposure;
- anything that can't be undone.
An agent may prepare them; a person decides them.
Questions to ask before your first agent.
Which work takes up most of your team's week?
Where does information get typed in twice?
What must always have a person's yes?
How would you know it was working?