How to Choose an AI Consultant for Small Business

Choosing an AI consultant or implementer is hard when you are not already an AI expert. The field is full of people who can make impressive demos but struggle to deliver real results. This guide helps you ask the right questions and spot the red flags before you commit.

What to Ask Before You Hire

Ask every candidate these questions and watch how they answer as much as what they say.

"Walk me through a project that did not work." A good consultant can describe a failed project honestly, what they learned, and how they changed their approach. If someone has never had a project fail, they are either lying or have not done enough real work.

"How do you handle scope changes?" The wrong answer is "we handle whatever comes up." The right answer involves a change control process, a written scope document, and a conversation before extra work begins.

"What happens after the automation goes live?" You want to hear about documentation, training, and a maintenance plan. The wrong answer is vague or assumes the client will figure it out.

"Who actually builds the automation?" Some consultants sell and then subcontract the build. That is not always bad, but you should know and have confidence in the actual builders.

"Can you show me something similar to my workflow that you have actually built?" Ask for a live demo or a reference from a business with a similar process. If they cannot show you anything, that is a red flag.

Red Flags to Watch

These are the signals that a consultant may not be the right fit for a small business:

The pitch is a platform demo, not a conversation about your problems. If someone shows up with slides but has not asked you a single question about your business, they are selling a product, not solving your problem.

They guarantee outcomes. No one can guarantee specific time savings or error reduction without knowing your data and process. Anyone who guarantees numbers upfront is either naive or lying.

They do not ask about your data. If someone wants to build without understanding where your data lives, how clean it is, and what access you have, they will hit blockers.

Their smallest engagement is a 6-figure project. Small businesses should be able to start with a narrow pilot for a few thousand dollars. If the minimum is high, either the consultant is not right for small businesses, or they are padding the scope.

They use a lot of jargon and do not explain things in plain English. You should be able to understand what they are proposing. If the explanation is all buzzwords, they may not understand it themselves.

How to Structure a First Engagement

A good first engagement is narrow, bounded, and produces a learnable result. The goal is to validate that the consultant understands your problem and can deliver before you commit to a larger engagement.

Recommended structure:
- Start with a single, well-defined workflow (not "improve our whole sales process")
- Define a clear success metric before you start (not "make it better")
- Set a fixed budget or capped hours, not open-ended time and materials
- Include a checkpoint review at the halfway point
- Require documentation and team training as part of the deliverable

What a good scope looks like: "Automate the follow-up sequence for new quote requests from our website form. The automation should draft an email within 5 minutes of form submission, route to the CRM, and alert [owner] for review before it is sent. Success is measured as: average response time under 2 hours, and zero emails sent without review."

A good consultant will welcome your questions. They will not make you feel ignorant for asking them. If a consultant makes you feel rushed or foolish for asking questions, that is information about how they will treat you as a client.