Pricing
What it costs.
Five stages. Each is a fixed price you see before it starts, with a stopping point after every one. No stage exists to sell you the next.
The payback promise
If it won't pay for itself inside a year, I'll tell you not to build it.
Your Scope report separates what the work costs you today from what automating it would actually save — review time and exceptions included. If the math doesn't clear, that's what the report says, and you've spent $2,500 instead of much more finding out.
Assessment
Free
A short questionnaire about how your business actually runs. I score every repetitive workflow you have and pick the one worth attacking first.
Deliverables
- A ranked shortlist of your own workflows
- A written case for the top one, with what it's costing you per year
- A slide deck you keep either way
Price
Free
Book a call, or don't.
Scope
Fully remote
You record yourself doing the work once. I watch the real thing — including the parts that go wrong — walk the open questions with you on a call, and test whether your data source can actually be automated.
Deliverables
- Your workflow documented properly
- What automating it would genuinely save — separated from what it costs you today
- A build range and a payback estimate
- A straight go or stop — and if it's a stop, where to look instead
Price
$2,500
Flat. Same for every client.
Build, or take the report and stop.
Proof
One week, fixed
I build the roughest thing that can process one real batch of your work, then run it alongside you for two live cycles and count every time it disagrees with you. No polish, no error handling, nobody but me can run it. Anyone can show you a prototype that runs; this tells you whether it agrees with the person currently doing the job, on your actual work. That's the difference between a demo and evidence.
Deliverables
- A working rough system, running on your real data
- A disagreement count against live batches — how often it got it right
- A costed list of exactly what production would take
Price
$7,500
Flat. Same for every client.
Go to production, or stop with the answer.
Production
Quoted only after Proof
Error handling, monitoring, whatever interface your team actually needs, every edge case Proof turned up, deployment and handover. Priced from that counted list — not from a guess, and not before.
Deliverables
- A system your team runs without me
- Documentation and handover
- Milestone payments: skeleton running, running in parallel, live
Price
Quoted from Proof's findings
Tiered must-have vs nice-to-have, so the budget sets the scope rather than the reverse.
Support
Monthly, cancel any time
A custom system needs a home. Monitoring, the day something it depends on changes without warning, model and API upkeep, small improvements as the work shifts, and me being reachable. Breakage and running costs are in the fee — you will never get a variable invoice for something you can't evaluate.
Deliverables
- Monitoring and fixes included
- Model and API costs included, to a stated fair-use ceiling
- Small iterations as the work changes
Price
From $1,000/mo
Scaled to the size of the system and what it touches. Quoted with your build, never after.
Month to month. Leave whenever.
What I won't do
I won't quote a build before I've watched your work.
Anyone who does is guessing, and you'll pay for the guess either way — in a padded price or in change orders.
I won't bill you by the hour.
An open meter puts the risk on you. Every stage is a fixed number agreed before it starts, and if it takes me longer than I expected, that's mine to absorb.
I won't credit one stage against the next.
Each stage has to be worth its own fee, because that's the only way "don't build this" stays a real answer.
Questions I get.
Why does Scope cost anything if the assessment was free?
The assessment ranks your workflows from the outside. Scope is me watching the actual work and testing whether your data can be automated at all. It's the difference between a hypothesis and a finding, and it's what makes a build price real.
Someone quoted me a few thousand for a working prototype. Why is Proof $7,500?
Because a prototype answers "can this be built," which was never really in doubt. Proof answers "does it agree with the person doing the job right now" — I run it against your real batches and count the disagreements. That number is what tells us whether production is worth paying for, and it's the one thing a demo can't give you.
Can I just pay for the build?
No. That's the version where I quote you a number I can't stand behind and one of us absorbs the difference. The stages exist so neither of us is guessing.
What if Scope says don't do it?
Then that's the report, and it ends with the next workflow worth looking at instead. You keep the documentation and the numbers, and they stay useful long after the decision — for training new staff, handing the work over cleanly, or showing a buyer how the business actually runs.
Why not credit the $2,500 toward the build?
Because then the report has a commercial reason to say yes. Keeping it standalone is what lets me tell you to walk away.
Is $2,500 refundable?
No. It's the stage that makes sure the rest is money well spent — a real scope, a real number, and a build shaped around how you actually work rather than how I assumed you do.
Who owns what you build?
You do. Code, prompts, documentation.
Every engagement starts the same way: a free assessment of where AI actually fits in how you work. You keep the results either way.
Find where AI fitsRather talk it through first? Book a call.