Most operations problems get counted in the wrong unit. A tool costs $25 a seat, so the tool looks cheap. The hours spent around it never reach an invoice.
Those hours are the leak. The Friday afternoon a project manager spends reconciling two systems that disagree. The approval chased over email because the portal cannot hold it. The client status update three people assemble by hand. Nothing errors, and all of it gets paid for every week, at salary rates.
This audit counts those hours, then prices the fix.
What I do
- Talk to the people who move the work, not only whoever chose the tools. The person reconciling a spreadsheet on Friday knows where the time goes, and usually nobody has asked them.
- Inventory the stack. Every tool in the workflow, the seats, the monthly cost, and what each one is really used for. That last column is where the surprises sit.
- Watch the expensive workflows run. A walkthrough or a screen recording of the few workflows that eat the most time, so the count comes from what happens rather than from what the process document says happens.
- Count the leak. Hours a week per workflow, each figure written next to the interview or recording it came from, so you can check every number yourself.
- Price the fix. What to change, what to keep, and a fixed price for anything worth building.
What you receive
A written report and a readout call, about two weeks after kickoff. The report is yours either way. If the fix is a build, the report already carries its fixed price, so the next conversation starts from a number rather than from another discovery call.
Sometimes the answer is the tool you already have
I publish comparisons of five tools against a custom build, and each one says when the tool is the right answer. The audit holds to the same rule. If a cheaper fix inside the tool you already pay for beats replacing it, the report says so on page one.
Why a fixed price
Because the usual alternative is a discovery phase billed by the hour that ends in a proposal. A fixed price means you know what the answer costs before you ask the question, and you decide on a build with the numbers in front of you.
What this is not
Not a code review. If the question is whether a codebase written largely with AI will hold up under diligence, that is the AI Code Confidence Review.
Not a sales pitch dressed as a diagnostic either. The report is written to be useful to someone who never speaks to me again.
Pairs with Operations Infrastructure when the fix is a build, and with Fractional AI Ops Lead when the system already exists and needs watching.