Accounts receivable management with AI: who still needs to pay and what you can do about it

Around a quarter of B2B invoices in Western Europe are paid after the due date, according to the Payment Practices Barometer for Western Europe 2026 from credit insurer Atradius. That is not a side issue. It is your working capital sitting in your customers' accounts.

Yet many business owners let debtor follow-up slip. Not out of reluctance, but because it is a hassle: running a report in Exact Online, going through the list, drafting emails, then again next week. With an AI assistant on your administration, that becomes a conversation of a few minutes. Below you can read how to request the ageing analysis in plain language, how the AI prepares a payment reminder for you and how to track the trend using DSO.

Request the ageing analysis in plain language

One question is enough: "Give me the ageing analysis of my debtors." The only preparation is a one-off connector between your own AI assistant, such as Claude or ChatGPT, and your administration via Exact AI Connect. The AI reads the ageing report from Exact Online and summarises it for you.

In the answer you see per customer how much is outstanding and how old those items are: just overdue, more than 30 days, more than 60 days or even more than 90 days. That way you see at a glance where the money is and where the risk lies. And because it is a conversation, you simply keep asking: which invoices sit behind that balance, when did the oldest one fall due, is this a pattern? The AI points you to the invoice in Exact Online with a direct link, so opening the source document takes one click.

How such a conversation sounds in practice

A realistic flow, without you having opened a single report yourself.

You: "Who still has to pay? Sort by outstanding amount."

The AI puts your largest customer at the top, with three outstanding invoices, one of which is well past the due date. Below that is a wholesale customer with one invoice that fell due last week. Per item you see the invoice number, amount and the number of days since the due date.

You: "Is that usual for that wholesale customer, or do they normally pay on time?"

The AI looks at the customer's earlier invoices and reports that these have generally been paid neatly within the term. One late invoice is then more likely a forgotten payment than a problem case; a friendly reminder will probably do.

That is the difference from a static report. You do not just get the list, but also the context to decide what to do with it.

Let the AI write the reminder, send it yourself

Anyone searching for an automatic payment reminder for Exact Online often hopes for a button that handles everything. The honest story: the AI never sends an email itself. What it does do is write a draft based on your outstanding items that you can use straight away. Ask, for example: "Write a friendly payment reminder for all invoices that are more than 30 days outstanding." You get a draft per customer with the right invoice numbers, amounts and due dates, in the tone you choose: friendly for someone who is late once, more businesslike for someone who structurally stretches things out.

After that you read it over, change whatever you want to change and send the email from your own mailbox. That last step deliberately stays with you. It is your customer relationship, and a reminder that lands the wrong way costs more than it brings in. We believe sending should remain people's work.

Measure your progress with DSO

Individual reminders stay symptom relief if you do not know whether the underlying picture is improving. That is what DSO measures. Credit insurer Allianz Trade describes DSO (Days Sales Outstanding) as the average number of days invoices remain open before the customer pays. The calculation: DSO = (accounts receivable balance / total revenue) × the number of days in the period.

A reliable Dutch benchmark honestly does not exist; the figures in circulation are dated or not geared to SMEs. So compare your DSO with your own history and with the payment term you agree. If you use 30 days and your average sits well above that, you know enough.

You simply ask your assistant this too: "How many days do my sales invoices remain open on average, and how is that developing this year?" For a trend analysis like this over longer periods, the connector uses the analysis function; that is included in the trial period and in the analytics variant of the subscription.

Turn it into a short weekly routine

Accounts receivable management wins or loses on consistency. Set aside a fixed moment for it every week, Monday morning over coffee works well, and keep it small:

• Request the ageing analysis and scan what has newly fallen due since last week
• Have draft reminders written for the items that are ready for one
• Send the emails and note the payment arrangements that customers make
• Watch the outliers: a customer with a large balance deserves a phone call, not a third email

It does not have to take much time. If your sales team works in a CRM, make it visible there too who still has to pay; you can read how to approach that in our article about payment status in your CRM.

Example prompts for your accounts receivable management

Copy these questions into your connected AI assistant and adapt them to your own receivables:

Give me the ageing analysis of my receivables.

Who still has to pay? Sort by outstanding amount.

Write a friendly payment reminder for all invoices that have been open for more than 30 days.

How many days do my sales invoices remain open on average, and how is that developing this year?

Want to try it yourself? Start for free with Exact AI Connect →