How it works
The version that fails costs six months and produces a tool nobody opens.
This is what the other version asks of you, and what you are left holding.
What we need from you
Four things, and we say up front what each costs in hours.
If a deployment needs more of your people than this, it is too big for a first engagement and we would rather narrow it.
- The person who does the work
- Not their manager. They are in the room for the build, because they are the one who keeps it running.
- Access to the real thing
- The live system, the actual messy spreadsheet. Anything built against a tidy sample fails its first week in production.
- A baseline you agree with
- Before we change anything we measure what the work costs now, in hours and errors. You sign off on that number.
- Someone who can say yes
- Deployments stall on approval, not on engineering.
What you keep
When we leave, the thing that runs your workflow is yours.
No licence, no seat count, no dependency on us picking up the phone. That is a deliberate limit on what we are willing to build.
| You get | Detail |
|---|---|
| The working system | In your own accounts and your own infrastructure, not ours. |
| The build history | Why each step is shaped the way it is, including what we tried and rejected. |
| The measurement | Before and after on the same basis, so the result is yours to defend. |
| The next candidate | A ranked list of the other workflows we saw while we were in there. |
What we turn down
Naming these now is cheaper than finding them in month three.
We do not take work where the deliverable is a strategy document. We do not automate a process whose owner has not agreed it needs changing. We do not replace a team, and we will say so if that is the unstated brief.
We also will not start where the underlying process is broken in a way AI does not fix. Automating something nobody trusts produces faster output that nobody trusts.
Cost
Priced against the baseline, not against a rate card.
The fee is set per engagement once we have measured what the workflow costs you today, because that number is what the work has to beat.
Which means we can tell you what it costs on the first call, once we know which workflow you have in mind. We would rather do that than publish a range and spend the call defending it.