The short answer

Look for someone with operating experience who can find the first process worth fixing, build it into your workflow, and help the team use it. An independent implementation strategist or small hands-on firm can fit that job.

The automation builder. Usually a freelancer. Connects the tools you already have, wires up a few workflows, hands you something that runs. Good when you already know what you want built. Not good when you do not, because they will build whatever you ask for, including the wrong thing.

The boutique agency. Small team, packaged offers. A chatbot. A content engine. An onboarding flow. Good when your need matches something they have shipped before. Less good when your problem is unusual, because the package is the product and your business gets shaped to fit it.

The trainer. Comes in, teaches your team prompting and tools, leaves. Genuinely valuable, and almost always bought too early. Training a team on tools before you have fixed the process underneath is how you get a staff that is very good at doing the old thing slightly faster.

The developer. Builds custom software. Hire one when you have a specific product to build and you already know the spec. Do not hire one to figure out what your business needs. That is a different job, and paying build rates for discovery is the most expensive way to find out you did not need the build.

The strategist who implements. Watches how work moves through your business, picks the order, builds the first thing, and stays through adoption. This is the one most small businesses are looking for and the hardest one to identify from a website, because everyone else claims the same territory.

They hire for the model.

Business owners come in asking which AI to use, and that is the smallest part of the problem. The proportion is roughly ten percent model, twenty percent technology and data, seventy percent people and process. I wrote out where that breakdown comes from and why it holds in What Is the 10-20-70 Rule for AI?.

Read the split again and the hiring question answers itself. If seventy percent of the outcome is people and process, hire for the seventy. Almost nobody does. They hire for the ten, get a working tool nobody uses, and conclude AI does not work for a business like theirs.

Order.

Most small-business AI projects stall because the tool got installed on top of a process nobody had cleaned up, in a business where nobody had decided what came first. Foundation before growth. Operations before the exciting use case. That is the whole thesis behind The Sequence Model, and it is the first thing I look at on any engagement, before a single tool gets named.

So you are hiring for a different skill than tool knowledge. You want someone who can walk into your business, watch how work moves, and tell you which single thing to fix first. Then build it. Then stay.

“What would you do first, and why that?” A good answer names a specific process in your business and a reason. A weak answer names a tool. If they cannot answer before they have looked at your operation, that is a fine answer too, as long as they say so instead of pitching.

“Have you run a business, or advised one?” Both are legitimate. They are not the same. Someone who has sat in the meeting where a team quietly decides not to use the new thing builds differently than someone who has only presented at that meeting.

“Who owns this after you leave?” If the answer is you, ask how. If the answer is a document, keep looking. A real handoff names a person on your team, a way to tell whether it is working, and what happens when it breaks.

I put the longer version of this vetting conversation, including the red flags, in How Do I Choose an AI Consultant?.

It varies more than any other professional service I know of, because the category is young and nobody has standardized what they are selling. Hourly, project, and monthly retainer are all common, and the same scope can be quoted three different ways by three people in the same week. I broke down the real ranges and what drives them in How Much Does an AI Consultant Cost?.

The useful filter is what the price is attached to. A deliverable, or an outcome. Paying for a strategy document is the easiest money to waste in this category.

That is often the right call. There is a real amount of ground you can cover yourself first, and doing it makes any engagement afterward cheaper and faster, because you show up knowing where the time goes. I laid out what that looks like in How Can AI Be Used in Small Businesses?.

Start there. Come back when you hit the wall where the fix stops being a tool and starts being a decision about how your business runs. That wall is the moment to hire.

Ask who will be there in month three.

The one who can tell you what to do first, do it with you, and hand it over in a way that survives their leaving. If you want to see how I structure that, it is on the offerings page. Or just book time and we will look at your actual business.

Book a 30-Minute Strategy Call

If seventy percent of the outcome is people and process, hire for the seventy. Almost nobody does.

Should I hire an AI consultant or train someone on my team?

Do both, in that order. Training a team before the process is fixed makes people faster at the old workflow. Bring in outside help to decide what changes and build the first version, then train your team on the thing that now exists. The internal owner matters enormously, but they need something to own first.

How much does an AI consultant cost for a small business?

There is no standard rate. The category is young, and hourly, project, and retainer pricing all coexist for the same scope. What matters more than the number is whether you are paying for a deliverable or for an outcome. The full breakdown is in How Much Does an AI Consultant Cost?.

What is the best AI to help me with my business?

Almost certainly one you already pay for. Most small businesses have unused AI features sitting inside their existing email, documents, CRM, and accounting software. Buying another subscription before you have used those is the most common early mistake.

How do I use AI to help a small business without hiring anyone?

Pick the single task that eats the most time and is the least interesting, and automate only that one. Do not start with the exciting use case. One boring process, done properly, teaches you more about where AI fits in your business than a dozen experiments.

What is the difference between an AI consultant and an AI developer?

A developer builds what you specify. A consultant helps you work out what to specify. If you already have a clear spec, hire the developer. If you are still deciding what the business actually needs, hiring a developer first means paying build rates for discovery.

Want help finding your first AI win?

Let’s talk about what you need and how we could work together.

Book a 30-Minute Strategy Call