There's a particular kind of freelancer forum thread that always catches our attention: the one where someone quietly asks "does anyone actually use AI for their invoicing?" and then twenty replies come in from people who've apparently been doing it for months without mentioning it. We've been following a few Irish freelancer communities and co-working Slack groups over the past while, and a handful of tools keep coming up. Not the obvious ones either.
The tools people are actually talking about
1. Notion AI for client-facing documentation
Notion has had an AI layer built in for some time now, but freelancers in particular seem to be finding it useful for drafting proposals and project briefs. The appeal is that you're working inside a document you'd be writing anyway, so the AI assistance feels natural rather than bolted on. A few people in Irish remote-work groups mentioned using it to structure project scopes before sending them to clients, which honestly reminded us of the approach we wrote about in how to use AI tools to write a proper spec for a builder before you sign anything. Same principle, different industry.
2. Copilot in Microsoft 365 for admin-heavy weeks
Yes, this one is more well-known, but the specific use case coming up in Irish freelancer circles is email triage and meeting summaries for people who do a lot of client calls. If you're already paying for a Microsoft 365 subscription, it costs less than you might expect to add Copilot. Worth checking the current pricing in euro before committing, as it does shift.
3. Xero with its built-in AI features for VAT
This is the one that comes up most often in threads specifically about Irish freelancing. Xero has quietly added a lot of AI-assisted categorisation and reconciliation features, and for anyone dealing with VAT returns on the Irish system, having software that at least attempts to categorise your expenses correctly saves a surprising amount of time. It's not perfect and you still need to check everything, but several people mentioned it catching things they'd have missed. If you've had mixed experiences letting AI near your invoicing before, it's worth reading about the mistakes people have made when letting AI write their business invoices before you hand anything over to an automated system.
4. Otter.ai for transcribing client calls
Particularly popular with freelancers doing discovery calls or consultancy work. You record the call, Otter produces a transcript, and you can then ask an AI to summarise the action points. It's not flawless with strong regional accents, a few Irish users noted, but it's accurate enough to be genuinely useful. The free tier has limits but covers a reasonable number of hours per month.
5. Claude for writing that needs to not sound like AI
This one comes up repeatedly in threads where freelancers are being honest about using AI to help with client-facing writing. The general consensus seems to be that Claude produces output that needs less editing to sound like a human wrote it. For copywriters and content people especially, that's the whole game. It handles nuance better than earlier models, and several people in these communities use it specifically for anything that goes directly to an end client.
One thing almost everyone in these conversations agrees on: whichever tool you use, the time you save on the task itself tends to get eaten up fairly quickly by learning how to prompt it properly. That learning curve is real, and it's worth factoring it in before you commit to a new subscription.