Start with one workflow. Stay for the system.

The first piece of work is one workflow at a fixed fee, quoted after a walk-through and credited toward the partnership if we keep going. I do the mapping, the design and the code myself.

Book a walk-through

The first workflow

Map
Step 1
Two or three working sessions with the people who actually run the process. Not their manager's version of it. We write down every step, every workaround and every "oh, and on Fridays I also check this". Then I design the app you'd build if you were starting fresh.
You give
About three hours from the people who do the work
You get
A process map and an agreed app design, yours even if you stop here
Build
Step 2
A working app connected to your real systems. Your team runs it next to the old way for a few days and tells me where it's wrong. That's the checkpoint: if it isn't earning its keep by the end of week two, we stop.
You give
Honest feedback, including "this is useless"
You get
A working app on your data, and a clear go or stop
Harden
Step 3
The boring bits that decide whether people trust it. Edge cases your team found, sign-in and permissions, writing results back into your tools, and alerts that tell us when an input changes shape.
You give
A sign-off from whoever owns security
You get
SSO, roles, an audit log and monitoring
Launch
Step 4
It goes live in production and the old spreadsheet gets retired, properly, so nobody keeps updating both. Then you decide what happens next.
You give
A decision: keep going as partners, or take the keys
You get
The app in production, the repo, a runbook and a recorded walkthrough

After the first workflow, I stay on.

There's usually a queue of workflows behind the first one. The partnership works through it without starting from scratch each time.

A roadmap we review every month. What's next, why it's next, and what it's worth.
Small slices, shipped one at a time. Each one live before the next one starts.
Your team gets stronger. I pair with your people so they can run and change what's built.
Month to month. A month's notice to stop, and you keep everything.

What counts as one workflow

One thing that starts it, one result at the end, and no more than three systems in between. Bigger than that and it goes on the partnership roadmap as several slices, each one shipped on its own.

One workflow

Flag supplier shortfalls before the buy deadline
Prep every sales call from CRM history
Close the books each month

Too big for one

Replace the ERP
Automate all of customer success
An AI assistant for everything

Questions people ask

Our data is a mess. Is that a problem?

Realistically, everyone's is. Week one is partly about finding out which mess matters. The app is built to cope with the data you have, not the data you wish you had.

Do we need engineers on our side?

No. If you have them, I'll work in your repo and your review process. If you don't, the partnership or a keep-it-running plan covers the upkeep.

Which AI models do you use?

Whichever does the job best for each step, and plenty of steps don't need a model at all. They can be swapped as better ones ship, without a rebuild.

Who pays for hosting and AI usage?

You do, directly. The app runs in your own cloud and AI accounts, so those providers bill you, and you'll see the costs before I do. I'll give you a monthly estimate in the scope and set spending alerts so there are no surprises.

Why only a few teams at a time?

Because I do the work myself. A few partnerships is how many I can do properly. If I'm full, I'll tell you on the call when I'll have room.

Start with thirty minutes.

Walk me through the workflow. You'll have a one-page scope in two working days.

Form not loading? Open the booking page

The first workflow

Map

Step 1

Two or three working sessions with the people who actually run the process. Not their manager's version of it. We write down every step, every workaround and every "oh, and on Fridays I also check this". Then I design the app you'd build if you were starting fresh.

You give: About three hours from the people who do the work
You get: A process map and an agreed app design, yours even if you stop here

Build

Step 2

A working app connected to your real systems. Your team runs it next to the old way for a few days and tells me where it's wrong. That's the checkpoint: if it isn't earning its keep by the end of week two, we stop.

You give: Honest feedback, including "this is useless"
You get: A working app on your data, and a clear go or stop

Harden

Step 3

The boring bits that decide whether people trust it. Edge cases your team found, sign-in and permissions, writing results back into your tools, and alerts that tell us when an input changes shape.

You give: A sign-off from whoever owns security
You get: SSO, roles, an audit log and monitoring

Launch

Step 4

It goes live in production and the old spreadsheet gets retired, properly, so nobody keeps updating both. Then you decide what happens next.

You give: A decision: keep going as partners, or take the keys
You get: The app in production, the repo, a runbook and a recorded walkthrough