AI for business, where it actually helps

Most people have tried ChatGPT by now. Plenty were impressed, and then it stopped there. That’s not strange. AI writing a poem about your company is fun, but it doesn’t save any time on Monday morning.

The value is in the everyday work. In the emails that need sorting, the questions that keep coming back and the texts that get written from scratch every time. That’s where I start.

Four places where AI usually helps

Customer questions. An assistant that answers the common questions, about opening hours, deliveries or how something works, and hands the rest to you. The customer gets an answer right away and you don’t answer the same thing for the tenth time.

Incoming email. AI reads, sorts and forwards to the right person. The urgent ones end up on top and nothing sits forgotten in a shared inbox.

Quotes and texts. A first draft of the quote, the product description or the reply to a customer in seconds. You read, adjust and send. You always have the final word.

Summaries. Long documents, meeting notes and email threads turned into short summaries with what actually needs doing.

None of this is science fiction. These are everyday tools, and once they’re in place you notice them in your calendar right away.

And where it doesn’t

AI is good at text, patterns and repetition. It’s worse at knowing when it’s wrong. That’s why I never build something where the AI makes decisions that cost money or customers without a person looking first. It writes the draft, you approve it. It sorts the email, you answer the important ones.

Sometimes the answer is that you don’t need AI at all. Some things that look like AI problems are better solved with plain automation, like a connection between two systems. When that’s the case, I’ll say so.

How to get started

A call. You tell me what a normal week looks like and where the time goes. No obligation, and in plain language.

A concrete proposal. I point out one or two places where AI would help most, and what it would take to build it.

A first build. We start small, with what gives the most for the least. When it works and you see the difference, we take the next one.

Want to know where it would help in your business? Get in touch and we’ll look at it together.

Your data

That’s a good question, and it should be asked before anything gets built. I choose services with clear terms on where data is stored and how long it’s kept, and we decide together what the AI gets to see. Social Security numbers, sensitive customer details and anything the task doesn’t need stay out. The AI should do the job, not collect information.

Advice and building in one

With most providers you get either a consultant who draws up a plan or a developer who builds what you ask for. I do both. I’ve worked with the web and systems for seventeen years, and with AI since it became genuinely useful. That means the proposal you get is something that can actually be built, and the person building it understands why.

If you want to read more first, there’s a longer guide in the article on where to start with AI in a small business, and one example of a quick win in the article on speed to lead.

Updated 29 September 2026