Consulting

AI consultants for small business. Most of what gets sold as AI is a process problem.

You're past the point where you can hold the whole business in your head. You're not yet holding systems that do it for you. That gap is where the week goes. The quote nobody chased. The job that got scheduled twice. The Friday afternoon spent rebuilding a number that already existed in two other places.

Most ai consultants for small business will look at that and quote you an AI agent. The ones worth hiring look at it and tell you which half of it isn't an AI problem at all.

What most AI consultants for small business get wrong

The sequence. That's the whole difference, and it's most of what separates an engagement that sticks from one that gets quietly switched off in month four.

A consultant who starts with the technology will find somewhere to put the technology, because that's what's on the price list. Ask for AI and you'll be sold AI. Ask for agents and you'll be sold agents. Nobody talks themselves out of an invoice.

But a business that can't say who owns a lead after 6pm doesn't have an AI problem. It has an ownership problem. Automate it and the lead still gets dropped, just faster and with better formatting.

So the first question is never where the AI goes. It's what's actually broken, and whether one person with a checklist would fix it cheaper. Sometimes the answer is the checklist. Sometimes it's a tool you already pay for and never configured. Sometimes it genuinely is AI, and then it's worth building properly instead of bolting on.

The consulting market has mostly split into two camps: vertical specialists who know your industry and sell one fixed answer, and horizontal generalists who know the tools and don't know your business. We've written about why neither camp asks the sequencing question. The short version is that asking it puts the sale at risk.

What the engagement actually is

Two weeks to findings. Not two quarters, and no discovery phase that bills while it decides what to look at.

Week one is mapping. We sit with you and with whoever actually does the work, and we follow a lead from the moment it arrives to the moment it's invoiced. Not the version on the org chart. The version where the office manager texts the tech directly because the scheduling app is slower than their thumbs.

Week two is findings. Every leak we found, what it costs you per month in your numbers rather than benchmark ones, and what fixes it, ranked by payback instead of by how interesting it would be to build. It runs against the same seven areas laid out on the assessment framework, applied to your operation. If missed calls are the leak you already suspect, the missed call cost calculator puts a monthly figure on that one before you spend anything.

What you're holding at the end is a document, typically 8 to 15 pages. Not a login. Not a dependency. It names the leaks, the order to fix them in, the tools that fit the stack you already own, and what each step costs. Build it with us, hand it to the person who already does your IT, or sit on it for a quarter. It's written to be handed over on purpose, because a recommendation you can only execute by rehiring the consultant who wrote it isn't a recommendation.

When AI isn't the answer, you get told that

That's an easy sentence to write and a harder one to mean, so here is what it looks like in practice.

A business comes in wanting an AI receptionist. The actual finding is that three people can book on the calendar and none of them can see the other two. Put an AI receptionist on top of that and it books the same slot three times, faster. The fix is a shared calendar and one rule about who owns it. That's a Tuesday, not an engagement.

The diagnostic we use for this is written up in full in That's Not an AI Problem. The compressed version: if the process would still be broken with an extra employee doing it by hand, AI won't fix it either. It'll run the broken thing at scale and send you a bill for the privilege.

You'll hear that on the first call, before there's an invoice in play. A consultant who won't tell you not to buy isn't giving you an assessment. That's a quote with extra steps.

Who this isn't for

Six ways to get told no by a business that would rather have your money.

Under $1M in revenue. The assessment is hard to justify at that size and the operational complexity usually isn't there yet. Fix the sales side first, come back after.
Solo operators and pre-revenue founders. There's no process between people to map when there's one person. You don't need a consultant, you need a customer.
Anyone shopping for an AI cofounder. If what you want is software that runs the company, that's a different vendor, and probably a different decade.
Over 50 employees. At that size you need procurement, an IT security review, and three approvals per decision. We're the wrong shape for that and you'd feel it by week two. We'll refer you out.
Businesses in genuine crisis. Cash flow, legal, an ownership fight. Systems work doesn't fix fundamental business health, and selling it to you mid-crisis would be taking money at the worst possible moment.
"Automate what we do now, don't change anything." The work requires something to change. If nothing can, automation just casts the current cost in concrete and makes it harder to undo later.

What operators ask before they hire

These are the six that come up on nearly every first call. Answered here so you can disqualify us without spending 30 minutes on it.

01 What does this cost?

The assessment is a fixed fee between $1,000 and $3,000, quoted before anything starts, and it's the same number whether the findings turn out interesting or boring. Build work is scoped separately, after you've read the findings, so you're never approving a build budget before you know what needs building. Nothing is on retainer and nothing renews on its own. If you want a firmer number before you talk to anyone, say so on the form and you'll get one in the reply.

02 How long before anything actually changes?

Findings land one to two weeks after kickoff. The first fixes are usually live in the two to three weeks after that, because the first fixes are almost always small: a shared intake inbox, a missed-call text-back, one rule about who owns a lead after 6pm. The larger builds run three to four weeks each and only start once you've approved them.

03 Are you going to sell me AI I don't need?

The assessment is priced as a standalone piece of work and it pays for itself whether or not you buy anything else. That's deliberate. It means the recommendation can be a $40 tool you already own, a process change with no software in it at all, or nothing, without anyone here losing money on the honesty. Process problems get found before automation opportunities because that's the order the assessment runs in, not because it's the answer you were hoping for.

04 Do you understand my industry, or will I be explaining it for three weeks?

Contractors and local services, law firms, real estate offices, agencies, e-commerce, and small manufacturing shops are the six we work in most. The surface symptoms differ by trade. The underlying diagnosis rarely does, because lead intake, follow-up, handoffs, and reporting break the same six ways everywhere. You'll spend one session explaining your business, not three weeks.

05 What if the answer is that I don't need AI?

Then that's the answer, and you'll hear it on the first call rather than after an invoice. If the process would still be broken with an extra person doing it by hand, automation won't fix it. It'll just run the broken version faster and more expensively. Telling you that is the job.

06 What am I left holding if we stop working together?

A written findings document, typically 8 to 15 pages, naming your leaks, what each one costs you in a month, and the order to fix them in. Anything already built runs in accounts you own, on tools you pay for directly, documented well enough that your existing IT person or a new hire can pick it up. Nothing is hosted here and nothing is hostage.

Next step

Book the call.
Find out if there's anything here worth doing.

30 minutes with the person who'd run the assessment, not a sales rep working from a script. You'll describe how work moves through your business. We'll tell you where it's most likely leaking and whether any of it is worth automating.

30 minutes, on Zoom or the phone, whichever you prefer.
You leave with a short written summary either way.
If there's no clear opportunity, you'll be told that on the call. It happens often enough to be worth saying out loud.

Still reading rather than booking? The field reports are where the same thinking gets applied to specific businesses, with the numbers left in.