Map. Blueprint.
Then decide.
Useful work starts before the first interview. I collect what a form can answer, then use the sessions to go deeper. You talk to me.
The three steps
Map the operation
Org chart, written procedures, and the job journey first. Then small interviews with the people who do the work. One person in the room, maybe two. Process, not performance.
Hand you a blueprint
Hours first, dollars second. Every finding cites a person or a document you already had. Recoverable time, where that time should go, and what is worth building first.
Build what earned it
Finding the leak is different from fixing it. The AI Audit decides: which parts AI can take off the desk, and which it cannot. Implementation is a separate, fixed-scope engagement inside the tools you already run.
Each of the five, with the symptom, the cause and a test you can run this week: where a trade shop leaks hours.
A shop that does not yet know what it needs.
That is the buyer. Not a company with an AI strategy and a vendor shortlist; a company with crews on the road, a back office that has not kept up with the work it already won, and a nagging sense that some of the typing should not be happening. The AI Audit is how you find out which parts, what they cost you, and what is worth doing about them, before you buy anything.
If you want the general answer before the audit, it is written down: what AI can actually do for a shop like yours, with a page per trade.
Trade shops
HVAC, plumbing, electrical, roofing. Crews on the road, a back office that runs the jobs.
Builders and remodelers
Jobs with a takeoff, a schedule, subs, and a draw.
Property managers
Units, tenants, work orders, vendors, owner statements.
- You suspect AI could help somewhere and nobody has told you where, honestly, without selling you the answer.
- The same facts get typed twice: a quote in a spreadsheet, then again in the job system.
- You bought software and the team went back to the spreadsheet.
- Quoting or work orders are where the week stalls.
Sized by headcount. A 15 to 30 person shop is about two days of interviews; up to 60 is three or four; above that, a full week. If you are not sure the audit is the right size for you, that is what the 30-minute call is for, and I will say so if it is not.
A written blueprint. Yours whether or not you hire me to build.
Presented live, then handed over as a document. Six parts.
The map
How a job is actually delivered in your shop, from the lead to the last invoice, drawn from the people who do it, not from the org chart.
Findings with a source
Every finding cites an interview or a document you already had. Nothing in the blueprint asks you to take my word for it.
The hours, then the dollars
Recoverable hours per week, what those hours are worth, and a recoverability figure below 100%, because nothing recovers all of it.
Where the time should go
Saved hours are only worth something if they land somewhere. The blueprint says where.
What AI can take, and what it cannot
Which steps are lookups, handoffs and re-typing that a system can absorb, and which are judgement, pricing and the customer conversation, which stay with a person.
A priority list, with risks named
Short-term and longer-term, in the order the numbers support. What to do first, and what to leave alone.
Two weeks, start to presentation.
Sized by headcount, because headcount is interview count. The useful work starts before the first interview, which is why the first week is mostly yours and the second is mostly mine.
What the first week asks of you is four things and one name, published in full: the checklist I ask for before an audit.
You name one owner on your side. I collect the org chart, written procedures, tool access and the job journey, and build a first map from them. Headcount sets the interview count.
Each person gets a short prep note: the tools they use, and a typical week in percentages. I read those before we sit down, so the interview goes deeper than the form.
Small interviews, 60 to 90 minutes, one person in the room, two at most. Process, not performance. About two days of interviews for a 15 to 30 person shop; three to four days up to 60; a full week above that. Transcribed.
Hours first, dollars second, recoverability below 100%. Findings tied to a person or a document. Then the presentation, live, with the written copy, and one of two outcomes: an agreed first priority, or a written decision to stop.
What I need from you
One named owner. The org chart, written procedures and tool access before interview week. The named people available for 60 to 90 minutes each. The typical-week table read before we sit down. That is the whole list.
A fixed fee, set by headcount.
The AI Audit is $5,000, fixed, sized by how many people I need to interview. There is no hourly meter and no second invoice. If the recoverable savings in the blueprint do not cover the fee, you do not pay, and you keep the blueprint.
The guarantee is checked against one number: the recoverable hours figure in the document you are holding. Not a feeling about the presentation.
Build what earned it. Or take the blueprint elsewhere.
Finding the leak and fixing it are two decisions. If the blueprint earns a build, that is a separate, fixed-scope engagement, $5,000 to $10,000, inside the tools you already run: Jobber, Housecall Pro, ServiceTitan, Buildertrend, AppFolio, Buildium, QuickBooks, Sage, Acumatica, Bolt by ECI. Phased work across several systems starts from $15,000.
Or you take the blueprint to whoever you like, including your own team. It was written to be usable without me.
The kind of thing a blueprint earns.
Subcontract review from 2–3 hours to 15 minutes for a commercial masonry contractor.
A trade contractor, a stack of paperwork nobody had time to read properly, and a tool that reads it so a person can decide. The shape every good build has: AI takes the reading, the judgement stays with the principal.
AI Audit questions, plainly answered.
If the savings do not cover the fee, you do not pay.
I map the operation, put a dollar figure on the waste, and hand you a prioritised blueprint. Finding the leak is the engagement. Building the fix is a later decision.