How we work

Measure first. Build second. Prove it.

The same discipline used to design and run industrial plant: map the process, find where the time goes, build with checks in place, and measure the result the same way twice.

  1. Intake

    You fill in a short form and send a phone video of how things run today, so we start at the problems, not a sales pitch.

  2. Audit

    Your team logs how long 3 to 5 everyday jobs take for a week, and you walk us through them. That log becomes the baseline for every later claim.

  3. Roadmap

    A written plan: what to automate, the hours each change saves, what it costs and how fast it pays back.

  4. Build

    Built in your own accounts, from tested patterns, and checked against real examples of your work before anyone relies on it.

  5. Measure

    Thirty days after go-live the same people log the same jobs again. The before-and-after number is the result, whatever it says.

  6. Handover

    A one-page guide for each automation, a recorded walk-through, and the first month of the Care plan free.

The rules we work to

  • Automate the work, keep the judgement

    Claude drafts, checks and files. A person approves anything that goes to a customer.

  • Your accounts, your data

    Everything is built in accounts you own, in your business's name, with model training on your data switched off.

  • Checks, not good intentions

    Where a mistake would be costly, the check is built in and runs every time, not left to someone remembering.

  • A safe fallback

    Every automation can fall back to the way you do it now, so the work never stops.

Results log

Every client result, with how it was measured. Nothing here is an estimate.

26 October 2026
No client results published yet. The first measured client before-and-after is due in March 2027. Every engagement's result is posted here with its basis, including any that fall short.