Phase 01
We watch
We sit in while your team does the work as normal. Nothing stops or slows down while we learn it.
- We join a call while someone on your team does the task the way they always do
- We note the rules they keep in their head
- We work from the real job and its real exceptions, not a written spec
- We ask what the work costs you today in time and errors
- You come away with a date for when it goes live
So we automate the real job, not a tidy description of it.
Phase 02
We build
We fix the process, then automate it.
- We write the process down and agree the rules with you
- We fix what's broken before anything is automated, and the documented process is yours to keep
- We build it on our own platform, using APIs where a system has them and browser automation where it doesn't
- A person approves payments and anything else you say needs sign-off
- We test on your real data until it gets the known answers right
Automating a broken process just makes the mistakes faster. In 43 of our last 50 deals with a recorded buyer call, the process had a gap nobody had written down.
Phase 03
It runs
Every day, in the tools you already use.
- Results, approvals and alerts arrive in Slack, Teams, email or a spreadsheet
- A second AI check reviews every run against what it should have done
- Failed runs go to a queue our team owns, and you get the cause by the next working day
- When your business changes, we change the rules, on request
- If you need a dashboard on your own data, we build that too
Most of our work happens after go-live. That’s the part a typical outsourcer charges extra for.
What you can count on
- A go-live date after the first call
- A price per process before we start
- Every run checked and logged
- The documented process and your data stay yours