The Credit Control Nobody Owns

An invoice went out three weeks ago. It was due last Friday. You have thought about chasing it four separate times, decided each time that you would leave it another few days, and it is still sitting there. Nothing is broken. Nobody has forgotten. The chase simply keeps losing to something else, and it will keep losing, because the thing it keeps losing to is not workload.

What is credit control actually for?

Most people hear “credit control” and picture a person in a finance team with a phone and a spreadsheet. That is a job title, not the thing itself. The thing itself is a control, in the same sense as any other check that stops something going wrong: it exists to make sure money you have already earned actually turns up.

I spent years as an internal auditor testing whether controls ran. Not whether they existed on paper, which is easy, but whether they fired every time they were supposed to. There are only two ways it goes wrong. Either the control is not designed well enough to be worth testing at all, or it is designed properly and is not operating as intended.

Credit control in a small business usually manages both. Nobody sat down and designed it, so what stands in its place is a person deciding each time whether to chase. And the thing competing with that decision is not another task.

Why do I put off chasing invoices?

It is not time. Sending a reminder takes ninety seconds and everyone reading this has ninety seconds.

It is the relationship. Asking a client for money changes the temperature of the conversation, and if you like them, or you want the next project, or they have been slow before and you do not want to be the supplier who nags, you will find a reason to wait until Friday. Then it is Friday and you would rather not start the weekend on that note.

This is why credit control is the control most worth taking out of human hands. Not because people are careless, but because the decision itself carries a social cost, and a control that charges you something every time it fires is a control that will eventually stop firing.

Is my chasing process actually working?

The audit test for any control is short. Ask these about your own chasing and the answer will usually be uncomfortable.

Who owns it? Not who does it when they get round to it. Who is accountable if it does not happen. In most small businesses the honest answer is the owner, alongside forty other things, which is the same as nobody.

What triggers it? A date, or a feeling? If the trigger is “when I notice”, the control does not have a trigger. It has a mood.

What happens next, and who decides? If the escalation from reminder to formal notice is decided in the moment, it will be decided differently every time, and mostly it will be decided to wait.

Can you prove it ran? Pick a client who paid late six months ago. Can you say what was sent and when? If not, you cannot tell whether the process failed or was never applied.

How do I chase without the awkwardness?

Here is the move that does most of the work, and it costs nothing.

Put the chasing schedule in your terms, so the client reads it before the first invoice exists. Reminder on day 7. Second on day 14. Formal notice at day 30, at which point statutory interest applies under the Late Payment of Commercial Debts (Interest) Act 1998. Final notice at day 45.

Two things change once that is written down. The client is not surprised, so a reminder is not an accusation. And you are not deciding anything when it fires, because the decision was made months ago by both of you.

How do I automate invoice chasing?

The mechanics are unglamorous, which is the point. Your accounting system already knows which invoices are overdue and by how many days. Something needs to read that and act on it on a schedule.

In practice that means your accounts package, Xero for most small businesses, holding the invoice data, and an automation tool checking it daily and sending the right message at the right stage. We build these on n8n, self-hosted, because it means the whole thing runs on infrastructure you own rather than a subscription that can change its pricing or disappear.

Stage five matters more than it looks. The fastest way to lose the goodwill you were protecting is to chase a client who paid yesterday.

Two things stay human, deliberately. Anything above the final notice is a decision with commercial consequences and belongs to a person. So does any invoice the client has actively queried, because that is a scope conversation wearing a payment problem’s clothes, and automating it just annoys everybody. If you want the wording for each stage, the payment chaser sequence prompt drafts all four, and there is a fuller walk-through of the timing in our guide to automated late payment chasing.

Where does my data actually go?

Any chasing setup handles real client information: names, contacts, amounts owed, and a history of who pays late. That is commercially sensitive about your clients as well as about you, so it is worth being deliberate rather than accepting whatever a tool defaults to.

Four things to hold to. Send the system only what it needs, which for a reminder is an invoice number, an amount and a date, and never a note about the client’s reliability. If any step uses an AI model, use a paid business tier with training switched off and a data processing agreement in place, or self-host an open model. Keep the data in the United Kingdom or the European Union. And prefer a self-hosted automation tool, because then the ledger data never leaves infrastructure you control.

Most of this is moot if you keep AI out of the sending, which as noted above you should. A reminder needs a date and a template, not a language model.