

Most small business owners hear "AI" and assume it's expensive, complicated or meant for tech companies. Some of the hype deserves that skepticism. But a few uses are practical right now for a business with two or five people, as long as you're clear about what the software does and where a person stays in charge.
This post is the big picture. If you're looking at a website chatbot in particular, read our checklist for buying an AI chatbot.
What AI is, in plain language
The AI tools that matter for a local business are language tools. They read and write text, and increasingly speech, well enough to draft a reply, summarize a long email, sort messages by topic or answer a routine question from information you give them. They don't know your business unless you tell them, and they can be confidently wrong. That second point shapes everything below.
Where it helps today
Answering routine questions
Hours, service area, what you do and don't do, how to book. A chat assistant on your website or a voice assistant on your phone line can answer these at any hour from a list of facts you approve, and take a name and number for anything else. The useful version knows when to hand over: "I'll have someone from the team call you back."
Drafting, not sending
Quote follow-ups, replies to reviews, a social post about a finished job, a summary of yesterday's calls. AI is good at a first draft. A person reads it, fixes it and sends it. That turns a blank page into an edit, which is where most of the time goes. The human step is the point, not a formality.
Sorting what comes in
An assistant can read incoming emails and form messages and label them: new quote request, existing customer, supplier, spam. It suggests, and your team decides what happens next. Sorting by clear rules you set, such as "anything mentioning a leak goes to the top", is safer than asking software to judge which customers matter.
Answering the phone when you can't
Voice assistants can now pick up, answer common questions, take details and pass the call or a summary to you. Done properly, it tells callers at the start of the call that they're talking to an AI assistant and offers a way to reach a person. Done badly, it pretends to be human, and customers notice.
Where it falls short
- Prices and promises. Don't let software quote a firm price or commit to a date without a person checking.
- Anything sensitive. Complaints, refunds, safety issues and legal questions need a person.
- Facts it wasn't given. If the answer isn't in what you approved, the assistant should say so and pass it on, not guess.
- Your voice. Drafts tend to sound generic. Edit them until they sound like you.
How we set it up
Erica is the assistant we set up inside your workflows — it drafts, sorts and answers routine questions, and a person approves anything that matters. It says it's an AI assistant at the start of every conversation. We start with one job, usually after-hours questions or review replies, check the results with you for a few weeks and only then add the next one.
Questions to ask before you start
- Which task eats the most time each week, and is it the same steps every time?
- What facts would an assistant need, and who keeps them up to date?
- What should it never do without a person approving it?
- How will you know it's working: fewer missed calls, faster replies, fewer repeat questions?
If you're curious what this would look like for your business, book a free 30-minute call and we'll go through those questions with you.
We do this for clients: AI integration.