We've been curious for a while about how AI tools are actually landing with small business owners here, not the tech-forward ones who are always early to everything, but the people running ordinary businesses in ordinary towns. So we asked around. Three owners, three very different businesses, and some pretty honest answers.
The Galway café
Sinéad runs a café near the Spanish Arch that's been open about six years. She started using ChatGPT last spring, mostly because a friend suggested it for writing social media posts. Her verdict on that: fine, but she edits almost everything it produces before it goes up, because the tone tends to come out a bit generic. "It sounds like every other café in the country," she said. She still uses it, just with more rewriting than she expected.
What she's kept without hesitation is an AI scheduling tool that helps her build the weekly rota. She has eight part-time staff with varying availability, and doing it manually used to take her the best part of a Sunday evening. Now it takes maybe fifteen minutes. That's the one she'd tell other café owners to try first.
She abandoned a tool that was supposed to help with stock ordering and waste reduction. It needed more structured data input than she had time for, and after a few weeks of trying to make it work, she let it go. "It wasn't stupid, it just needed more from me than I could give it."
The Wexford electrical firm
Paul has run a small electrical contracting business in Enniscorthy for about twelve years, mostly domestic and light commercial work. He came to AI tools later than Sinéad, and more reluctantly. His accountant mentioned that some businesses were using AI to help write quotes and job descriptions, and he decided to try it.
He now uses an AI writing tool, again essentially a version of ChatGPT, to draft the text sections of his quotes and the covering emails that go with them. He was surprised how much time that saved, because he'd always found the writing part of admin the most annoying. The actual figures he still does himself in a spreadsheet, but having the words done for him has made the whole process feel lighter.
He also tried an AI-powered job management app that promised to handle scheduling, invoicing and customer reminders. He used it for about two months. The scheduling part was genuinely useful. The invoicing section had enough quirks with Irish VAT and RCT that he didn't trust it, and he went back to his existing system. Worth mentioning, because we've heard similar things from other tradespeople: the tools built primarily for the UK or US market sometimes have enough tax and compliance differences to make them more trouble than they're worth here.
If you're in the trades and thinking about using AI at the proposal or quoting stage, the same instinct applies to bigger projects too. Getting the written side of a quote right matters, and it's something we've written about for people dealing with builders as well, even if that's the client side of the conversation.
The Dublin accountancy practice
Fiona runs a small practice in Rathmines with three other accountants and a part-time admin. She was the most measured of the three, which probably makes sense given the job.
Her team uses an AI summarisation tool to process long documents, things like lease agreements or loan facility letters for clients who want a plain-English version of what they're signing. She's clear that they always check the output carefully, but it saves time on a task that used to eat into billable hours. She also uses an AI tool integrated into their practice management software for drafting routine client emails and letters, which she described as "perfectly adequate for the boring ones."
What she's more cautious about is anything touching actual financial data or tax calculations. She's watched a few demo tools make confident errors, and her view is that the liability question isn't resolved clearly enough for her to rely on AI outputs without significant manual checking. For now, AI is doing the lower-stakes writing and reading tasks, not the numbers.
She mentioned one abandoned experiment: a client-facing chatbot that was supposed to handle basic queries out of hours. It gave a couple of answers that were close enough to advice that it made her nervous, and she switched it off.
What we took from it
The pattern that came through clearly was this: all three have kept tools that save time on writing and scheduling, and all three have walked away from tools that either needed too much setup or wandered into areas where an error actually matters. Nobody is using AI to replace their own judgement. They're using it to avoid typing things they'd rather not type.
That's probably the most practical summary of where things are for most small business owners right now: useful in the margins, unproven at the centre.