Four things we ask for
- Volume — how many of these a week, and how that varies
- Current cost — who does it now, how long it takes them
- Thirty real examples, including the awkward ones
- The destination — where the output has to end up
With those, an estimate is a calculation. Without them it is a guess dressed up as a proposal.
Why the awkward examples matter most
Thirty clean examples tell us the easy case works, which we already assumed. The three difficult ones tell us where the ceiling is, and the ceiling is what determines the business case.
So we ask specifically for the ones that would confuse a new starter, not the ones you would show a customer.
The conversation about the destination
Whether your accounting system, CRM or job platform has a usable API moves the price more than any AI decision. We ask early because the answer sometimes changes the recommendation entirely.
If there is no API and no import route, we will say so and price the alternative honestly rather than discovering it in week six.
When we say do not build it
- The volume is too low to repay any build
- A validation rule or a required field would prevent the problem entirely
- Your existing software already does this and it is switched off
- The process is about to change, so anything built now is wasted
We would rather lose a project than deliver one that cannot pay for itself. It is also, straightforwardly, how referrals happen.
What you get from scoping
A written scope with what is in and out, a fixed price for phase one, a duration, the assumptions it rests on, and the risks with how each would be handled.
If you take that document to someone else, that is fine. A clear scope is worth having regardless of who builds it.