DoIneedanAIconsultant,orcanwedoitourselves?

Many small businesses can start using AI without a consultant. If you want your team drafting emails, summarising documents or brainstorming with a well-known AI assistant, you can do that yourselves with a few sensible rules. You need outside help when AI starts touching customer data, connecting to your systems, or changing how work moves through the business. That is where mistakes cost more and are harder to see.

When doing it yourself is fine

DIY works well when the stakes are low and a person checks everything before it goes anywhere. Some examples:

• Drafting first versions of emails, adverts, job descriptions or social posts that someone edits before sending.
• Summarising long public documents, reports or meeting notes that contain no personal data.
• Turning rough bullet points into a tidy proposal or policy draft.
• Using AI features already built into your email, office or accounting software, following the supplier’s guidance.
• Individuals learning what the tools can do, on business accounts, with clear rules about what not to paste in.

If this is where you are, the most useful thing you can do is write a one-page usage policy, pick an approved tool on a business plan, and give people time to practise. You do not need a project.

Signs you need help

Bring in someone with experience when one or more of these is true:

• Personal or confidential data is involved. Customer records, health information, financial details or anything under a confidentiality agreement.
• It needs to connect to your systems. Reading from your CRM, writing to your accounts, sending messages on your behalf.
• It will act without a person checking every output. Automations and agents that send, file or update things on their own.
• You have tried and stalled. A pilot that worked in a demo but never made it into daily use is common, and usually a scoping or process problem rather than a technology one.
• Staff are using AI anyway, without rules. This is more common than people admit, and it is better to shape it than ignore it.
• You cannot tell which tools are worth paying for. The market changes monthly, and supplier claims are hard to compare.

What a good consultant actually does

A good AI consultant spends more time on your business than on the technology. Expect them to:

1. Understand the work first. Sit with the people doing the task, see where time goes and where errors creep in.
2. Recommend the smallest useful step. Often that is a better use of tools you already have, not a new build.
3. Agree boundaries before connecting anything. Which data can be used, where it goes, where a person reviews, and who owns what. At 3rd Eye AI these are agreed in writing before any data source is connected; see responsible AI.
4. Build or configure, then hand over. You should own the prompts, workflows and documentation, and understand how they work.
5. Train your team. People need to know what the system does, where it is weak and when to question it. That is the point of team enablement.
6. Tell you when not to use AI. Some tasks are better left as they are, or fixed with a simpler process change.

Questions to ask any AI consultant

• What have you built yourself? Can I see it working?
• How will you decide which task to start with?
• What data will you need access to, and how will you protect it?
• Where will a person review the output?
• Who owns the work when you finish, and can we run it without you?
• How are scope, fees and timelines agreed, and what happens if the scope changes?
• What are the ongoing running costs once it is live?
• Which tasks would you advise us not to automate?

Good answers are specific and a little cautious. Be wary of anyone who promises large results before understanding your business, or who wants a long contract before a small pilot has shown value.

A middle route

You do not have to choose between doing everything yourselves and handing everything over. A common pattern is a short piece of AI strategy work to pick the right first task and set the rules, then your own team runs the day-to-day use while outside help builds only the parts that need connecting or careful handling. Over time, your team takes on more.

For what I have built myself, The Counsel is an open-source AI legal assistant for England and Wales, and it shows the kind of care around sources, limits and review that business tools need too.

If you are unsure which side of the line you are on, get in touch. A first conversation is enough to tell whether you need help at all, and if you do, the scope, fee and timeline are set out in a written proposal.