The cost of AI automation, up front.
Here's what working with me costs. An operations audit — one week, your top three leaks quantified in hours and dollars — is $4,500, fixed. A build sprint that ships one working system is fixed-scope from $45,000. A fractional AI partner — ongoing senior engineering, month to month — is from $8,500 a month. Those are the real numbers. The rest of this page covers what each one buys and what moves the number up or down.
Almost nobody who sells AI automation publishes prices. The quote shows up after the discovery call, shaped by what the salesperson thinks you'll pay. I'm one engineer, on purpose, and I'd rather you disqualify me from this page than in the third meeting. So: numbers first, reasoning second.
Three engagements. Three numbers.
Every engagement starts with the audit. It's the smallest check you can write me, and it exists so neither of us builds on a guess.
| Engagement | Price | What you get | When it fits |
|---|---|---|---|
| Operations audit | $4,500 — fixed, one week | Your top three leaks quantified in hours and dollars, plus a build roadmap with the ROI math on paper | You suspect the leak but haven't measured it. Every engagement starts here — money-back if it doesn't pay for itself. |
| Build sprint | Fixed scope, from $45,000 | One working system — agents, MCP servers, integrations — shipped, documented, and owned by you | The audit found a leak worth more than the build costs. The $4,500 is credited toward it. |
| Fractional AI partner | From $8,500/mo, month to month | A senior engineer inside your business, shipping from a ranked automation queue every month | There's more than one leak, and you want the accountability of a hire without the $200K seat. |
The full mechanics live on the operations audit page; the ongoing option is covered on the fractional AI partner page.
Five things move the number. None are mysteries.
“From $45,000” is honest, not cagey. Here's exactly what pushes a build up or down from there — the same five questions I work through when I scope one.
- Systems touched. A workflow that lives in one system costs less than one that crosses four. Every added system is another connection to build, test, and keep secure.
- Read-only vs. write access. An agent that reads your data and reports back is the cheap version. An agent that changes records needs permissions, logging, and a way to undo mistakes — the guardrails are most of the work.
- Edge-case volume. If nineteen cases out of twenty follow the rule, the build is quick. If every third order is a special case, the build is mostly about the exceptions — and the price reflects that.
- Who owns the judgment calls. A system that hands unclear cases to a human is simpler than one expected to decide alone. Where that line sits is a design decision, and it moves the number.
- Existing data quality. Clean data in one place keeps a build fast. Three spreadsheets that disagree with each other mean I build the cleanup before I build the automation.
Fixed scope, not hourly. On purpose.
Hourly billing hands you all the risk. The meter runs, the estimate drifts, and the person doing the work has no reason to finish fast. Fixed scope flips that: we agree on the number and the definition of done before work starts. If the build takes me longer than I scoped, that's my margin, not your invoice.
The audit carries the same logic one step further. If it doesn't surface savings worth more than its cost, you don't pay — and when a build follows, the $4,500 is credited toward it, so you never pay for the map twice. The figure on the roadmap isn't a pitch. It's the size of the leak, measured in your hours and your dollars.
When the cheap option is right — and when zero is.
If your problem is “when a form comes in, add a row to the spreadsheet and email the crew,” you don't need me. That's a two-app trigger — a no-code freelancer can usually build it in an afternoon. It's cheaper and faster, and I'll tell you so on the first call.
And sometimes the right spend is zero. If the workflow still changes every week, automate it later — you'd be pouring concrete on a road that's still moving. If nobody on your side owns the outcome, the system dies quietly no matter who builds it. If the leak is smaller than the fix, keep the human. Automation is a purchase, not a virtue.
If you want your own number before you talk to anyone — including me — run the ROI calculator. Price the leak before you price the fix.
Keep reading: What to automate first covers picking the build these numbers attach to, and the ROI calculator turns your own hours and wages into the dollar figure that decides it.