AI consulting
Find out what AI is actually worth to you before you spend a dollar building it
We map how work really moves through your business, then rank the automation opportunities by payback. You end up with a costed plan you own, whether you build it with us or not.
Built for
- Owners who keep hearing “you should use AI” and want a straight answer
- Contractors and trades running on spreadsheets plus one overworked office manager
- Professional services firms billing hours they can’t bill for admin
- Government contractors who need a defensible plan before a spend request
- Teams that bought AI tools last year and can’t tell if they worked
AI readiness assessments are a fixed fee of $4,500–$9,000, set by how many processes and systems are in scope and quoted before we start. If you build the first project with us, the assessment fee comes off that invoice. Ongoing advisory retainers run $2,500–$12,000/mo.
The problem
The problem isn’t a lack of AI tools. It’s not knowing where they’d pay off.
You’ve seen the demos. A chatbot writes an email, everyone claps, and nothing changes on Monday. Meanwhile the real bottleneck is that your dispatcher retypes every job into QuickBooks, or your proposals take four days because three people have to touch a Word file. Nobody’s selling a demo for that.
So the decision stalls. Buying the wrong thing costs money and burns your team’s patience, and once they decide “that AI stuff doesn’t work here,” the next attempt is twice as hard. What’s missing is a clear-eyed look at where your hours actually go and which of those hours a machine can take.
What it buys you
The outcomes we hold ourselves to.
- A ranked list, not a wish list
- Every opportunity gets an estimated hours-saved-per-month figure, a build cost, and a payback period. The list is sorted by payback, so the argument about what to do first is over in one meeting.
- You find out what to skip
- Half of what people want to automate shouldn’t be. We tell you which ideas fail on volume, data quality, or liability, and why, so you don’t spend $30k proving it yourself.
- A map of where the work goes
- One page per core process (intake to invoice, RFP to submission) showing every handoff, every system, and every place data gets retyped. Most clients find 2–4 retype points they didn’t know existed.
- Honest answers about your data
- We look at your actual Jobber, HubSpot, or SharePoint data and tell you what’s usable now, what needs cleanup, and roughly how much cleanup. No AI project survives contact with bad data, and it’s cheaper to know in week two.
- A build spec you can shop around
- The deliverable is vendor-neutral and yours. Take it to us, to your internal dev, or to three other shops for quotes. We’d rather win the build on the plan’s merits.
- A number for the board or the bank
- Fixed-fee, fixed-scope, with a written estimate of return. That’s a document a lender, a partner, or a contracting officer can read without a translator.
Deliverables
What you actually receive.
Written, handed over, and yours to keep, whether or not we work together again.
Process maps
- One diagram per core workflow, showing systems, handoffs, and retype points
- Time-per-step estimates gathered from the people who actually do the step
- A marked list of every place data crosses between tools by hand
- Volume figures: jobs per week, quotes per month, invoices per cycle
The opportunity register
- 8–15 scored opportunities, each with hours saved per month and a build range
- Payback period per item, sorted worst to best
- A “do not build” section with the reasoning written down
- Dependencies flagged (e.g. “this needs the ServiceTitan data cleaned first”)
AI readiness assessment
- Data inventory: what lives in QuickBooks, Google Drive, SharePoint, Postgres, and someone’s desktop
- Data quality findings with sample records, not vague warnings
- Integration check: which of your tools have usable APIs and which don’t
- Security and access review, including who can see what today
- For regulated work: where CUI sits and what NIST 800-171 or CMMC obligations touch the plan
The roadmap
- A 6–12 month sequence with phases, not a big-bang plan
- Build spec for the first project, detailed enough to quote from
- Total cost of ownership per item, including licenses and ongoing hosting
- A one-page summary written for someone who wasn’t in any of the meetings
Process
How the engagement runs.
No surprises, no scope drift you didn’t agree to. Each phase ends with something you can read or use.
- 01
Kickoff and scope
Ninety minutes with you and whoever knows where the bodies are buried. We agree on which two or three processes to look at and who we need time with. There’s no prep homework, and we come to you.
Week 1 - 02
Sit with the people doing the work
Forty-five minutes each with the dispatcher, the office manager, the estimator, the proposal lead. We watch the actual clicks. This is where the real findings come from, and it’s the part most consultants skip.
Weeks 1–2 - 03
Look at the data
Read-only access to your systems where possible, exports where not. We check volume, field completeness, duplicates, and whether the timestamps mean anything. You get the findings raw, before we spin them.
Week 2 - 04
Score and rank
Every opportunity gets sized against hours saved, build cost, risk, and dependencies. We build the register and the “don’t build this” list at the same time.
Week 3 - 05
Walk it through with you
Two hours, on site or on a call, going through the roadmap line by line. Push back hard — we revise once after the session and hand over the final documents.
Weeks 4–5
Examples
The kind of work this turns into.
- A 40-truck plumbing outfit that didn’t need AI first
- The assessment found the dispatcher retyping every ServiceTitan job into QuickBooks, about 9 hours a week. That’s an n8n sync, not an AI project. The AI work (call transcription and summarization) went to phase three, after the plumbing was fixed. Illustrative of how these usually go: the cheapest win is rarely the fanciest.
- An 8-person engineering firm’s proposal bottleneck
- Mapping showed the four-day proposal cycle was really three hours of writing and three days of waiting for one principal to review. No AI fixes a queue of one. The recommendation was a routing change plus a draft-assembly tool built from past proposals, at a fraction of the cost of what they’d been quoted.
- A gov contractor with a data problem, not an AI problem
- Twelve years of past performance narratives across three SharePoint sites and two file shares, with no consistent metadata. The readiness assessment put a number on cleanup and staged the knowledge base behind it, so the capture team knew what they were buying before they bought it.
- A services firm that had already spent $20k
- They had three AI pilots, none in production. The register showed why: each one automated a step that wasn’t the constraint. We killed two, rebuilt one against the actual bottleneck, and the roadmap paid for itself on the kills alone.
Questions
Before you ask.
Is this just a report I’ll never read?
It’s roughly 20–30 pages, and the first one is the only one you have to read: the ranked list with costs and payback. Everything behind it is the evidence for that page. We also hand you a build spec detailed enough to get competing quotes, which is the opposite of a shelf report.
Why not skip this and just build something?
You can, and for a single obvious workflow (“stop retyping jobs into QuickBooks”) that’s often the right call. Ask us and we’ll tell you. The assessment earns its money when you’ve got several candidate projects, messy data, or a team that’s already been burned once.
Our data is a mess. Do we need to clean it up first?
No. Assessing the mess is part of the job, and we’d rather see it as it is, because the duplicates and blank fields tell us what’s automatable today and what needs work first. Cleaning up before we look just hides the problem we’re being paid to find.
Will you just recommend that we hire you?
The register includes items we don’t want and items you should handle in-house, and it’s written vendor-neutral so you can shop it. Plenty of assessments end with “fix this in ServiceTitan’s settings, no build needed.” We’d rather be right and get the phone call in a year.
How much of my team’s time does this take?
About 45 minutes each from 4–6 people, one 90-minute kickoff, and a two-hour readout. We work around your schedule, including early mornings before crews roll. Nobody fills out a questionnaire.
We’re a government contractor. Can you handle our environment?
Yes. We work read-only wherever possible, we don’t move data off your systems without written approval, and the assessment flags where CUI lives and what NIST 800-171 or CMMC obligations touch each recommendation. If something can’t be done in your enclave, the report says so instead of pretending.
Other services
Business Automation
Your systems talk to each other. Quotes, invoices, and follow-ups happen without anyone retyping anything.
Custom AI Solutions
AI built on your own documents, your own history, and your own rules, so it gives real answers instead of confident guesses.
Government & Enterprise
AI and automation that survive an ATO conversation, built for CUI boundaries, audit trails, and the review board.
Let’s find the work worth automating.
A 30-minute discovery call. We map where your time actually goes and tell you what’s worth automating, and what isn’t. No deck, no pressure. You leave with a written summary either way.