The short version: AI can write you a three-email invoice chasing sequence in under ten minutes that gets clients paying faster, but only if you feed it your actual contract terms and your actual client history, not a generic prompt. Do it wrong and you’ll send an automated late-fee warning to your best client the week before their renewal, which is a real thing that happened to someone I know.
Why I started using AI for this at all
For years I chased late payments the way most small business owners do: badly, inconsistently, and about three weeks later than I should have. I’d remember a client owed me money when I was doing my quarterly accounts, not when the invoice went overdue. By then it was £4,000 or £5,000 sitting there, and the conversation had gone from “quick nudge” to “awkward phone call I’d been avoiding.”
The Federation of Small Businesses has said for years that late payment is one of the biggest single reasons small businesses fail, not lack of demand, not bad products, just cash sitting in other people’s accounts instead of theirs. A survey a few years back put the average small business owed somewhere north of £20,000 in late payments at any given time. That’s not a rounding error. That’s payroll.
So in early 2025 I built myself a proper chasing system using ChatGPT, and it changed how quickly I got paid. Not because AI is magic, but because it removed the emotional friction that stops most of us sending the email in the first place.
The three-stage sequence I use
Here’s the exact structure, and you can copy this today.
- Day 1 after due date: Friendly nudge, assumes it’s an oversight, no mention of consequences.
- Day 7 after due date: Firmer, references the invoice number and original terms, asks for a specific date.
- Day 14 after due date: Direct, mentions statutory interest (in the UK you’re entitled to charge 8% plus the Bank of England base rate under the Late Payment of Commercial Debts Act), and asks for payment within 48 hours or a call to discuss.
My prompt for the first one looks something like this: “Write a short, warm email chasing an overdue invoice. Invoice number 1042, due 30 days ago, amount £2,400, client name Sarah, we’ve worked together for 18 months, tone should be casual not corporate, assume it’s simply been missed, no threats.” That last bit matters. The default AI tone is stiff and slightly passive-aggressive unless you tell it not to be.
By the second email I change the instruction: “Same invoice, second reminder, one week later, no reply received, still friendly but more direct, ask for a specific payment date this week.” By the third, I tell it plainly to reference the legal right to statutory interest and to offer a call.
The bit nobody tells you: automation without judgement will cost you clients
Here’s the uncomfortable part. A consultant I know, running a six-figure agency, set up a fully automated chase sequence through her invoicing software with AI-generated copy plugged in at each stage. It worked brilliantly for small, one-off clients. Then it fired the day-14 letter, complete with the late-fee threat, straight to her single biggest retainer client, worth £4,000 a month, two weeks before that client’s contract was up for renewal. The client hadn’t ignored the invoice. Their finance department had a new approval system and the payment was sitting one signature away from going out.
The renewal conversation that followed was not fun. She kept the client, barely, and only because she rang the finance director personally within the hour and apologised.
The lesson: automate the drafting, not the sending, for anyone paying you more than about £500 a month, or anyone you’ve worked with for over a year. AI writes the email. You still read it before it goes, and you still decide whether this particular client gets the standard sequence or a phone call instead. The tool doesn’t know your client is going through a rough patch, or that their usual finance person is on maternity leave, or that they’ve paid you on time for three straight years and deserve fifteen extra minutes of patience.
Why “polite” isn’t the same as “effective”
Most advice on chasing invoices tells you to be endlessly patient and understanding. I think that’s partly why so many small businesses stay broke. Politeness with no deadline attached gets filed under “deal with later” in someone else’s inbox, same as it would in yours. The version that works states a specific date and a specific consequence, delivered without apology.
Compare “whenever you get a chance” with “please can this be settled by Friday 14 August, after which a late payment charge applies under our terms.” One is a suggestion. The other is a fact with a deadline. AI is good at holding that firmer tone for you when your instinct, like mine, is to soften every sentence because you don’t want to seem difficult.
There’s a parallel here with how brands like Mailchimp think about their own email cadence. If you look at the Mailchimp marketing strategy, their whole system runs on precise timing and clear next steps rather than vague friendliness, and that same discipline applies to a chasing sequence. A deadline with no ambiguity does more work than a paragraph of niceties.
Adding a cost to lateness they can see
One thing I added this year that changed behaviour more than any email wording: a simple line showing what the invoice is costing them in real terms, calculated automatically. If an invoice is 30 days overdue at 8% plus base rate, that’s a specific number, not an abstract threat. I built a small interest calculator into my invoice template so the client sees the running total of what late payment interest they’d owe if I chose to enforce it, updated by date. It’s the same principle behind the interactive calculators I’ve written about for turning visitors into leads: a specific number someone can see changes their behaviour faster than a paragraph of text ever will. Nobody wants to watch a number tick upward with their name attached to it.
What Alex Hormozi gets right about cash flow that most small business owners ignore
I’ve written before about the business lessons from Alex Hormozi, and one of his blunter points applies directly here: cash collected is the only cash that counts. Revenue on paper with a 60-day payment gap isn’t revenue, it’s a promise, and promises don’t cover payroll. Most small business owners treat invoicing as an admin task that happens after the “real” work is done. It isn’t. Getting paid on time is part of the work, and treating it as an afterthought is exactly why so many otherwise profitable businesses run out of cash.
The businesses that chase, consistently, and without embarrassment tend to have stronger cash reserves than businesses with better products and worse collection habits. I’ve seen both sides of that up close.
Setting this up in your own business this week
You don’t need software. You need a spreadsheet, a calendar reminder, and a chat window.
- List every invoice currently over 14 days late, with the amount and the date sent.
- Write one master prompt for each of the three stages, with your real tone instructions built in (mine says “British, direct, no corporate filler, no exclamation marks”).
- Set a recurring calendar reminder for day 1, day 7, and day 14 past due date for every new invoice you send.
- Before sending anything to a client worth more than 10% of your monthly revenue, read it yourself and decide whether the standard sequence fits, or whether this one needs a phone call instead.
- Track how many days you’re getting paid faster over a quarter. Mine went from an average of 34 days to 19 days once I stopped waiting to feel ready to send the email.
If you’d rather have someone set the whole system up, with the prompts, the templates, and the escalation rules built around your actual client list rather than a generic script, that’s exactly the kind of practical automation an AI consultant for small business should be doing for you in a single afternoon, not a six-week project.
The follow-up discipline most businesses never build
Companies like Ahrefs built enormous trust in a crowded market partly through relentless, unglamorous follow-through, the kind covered in the Ahrefs marketing strategy, where consistency mattered more than cleverness. Chasing invoices is the least glamorous part of running a small business, and it’s exactly the kind of task AI is suited to, because the problem was never knowing what to write. It was remembering to write it, and having the nerve to send it without three rounds of second-guessing. Take the emotion out of the drafting, keep the judgement in the sending, and you’ll get paid faster without becoming the person clients dread hearing from.
Frequently asked questions
Is it legal to charge interest on late invoices in the UK?
Yes. Under the Late Payment of Commercial Debts (Interest) Act, UK businesses can charge statutory interest of 8% plus the Bank of England base rate on invoices paid late by another business, along with a fixed compensation fee. You need to state this in your terms and conditions or invoice for it to be enforceable, so add it now if it isn’t there already.
Will using AI-written chase emails make me sound cold or robotic?
Only if you let the default tone through unedited. The generic AI output tends to be stiff and overly formal, so tell it explicitly how you talk, feed it real details about the client relationship, and read every draft before sending. The AI is a drafting tool, not a replacement for your own read of the situation.
Should I automate the sending as well as the writing?
For small, occasional clients, automated sending through your invoicing software is fine. For any client worth more than roughly 10% of your monthly revenue, or anyone on a long-term retainer, read the draft and decide yourself whether the standard sequence fits, because a badly timed automated late-fee threat to a good client can cost you the relationship.
How much faster will I get paid using this system?
Results vary by industry, but in my own business, moving from irregular manual chasing to a structured three-stage sequence starting the day after the due date brought my average payment time down from around 34 days to 19 days within a single quarter.
Related reading: Instagram Marketing Strategy: How They Built a Brand That Wins and Depop Marketing Strategy: How They Built a Brand That Wins.
If you want the full breakdown, here is everything I know about AI marketing.
