Customer silence not always refusal to pay

In business-to-business relationships, a customer’s lack of response to an invoice does not necessarily mean they refuse to pay, according to a report from a small and medium-sized enterprise (SME) in trade. Sophie, a professional at the accounts receivable department of an SME, found the same customer at the top of her list of overdue invoices, with no response to three reminders.

Common causes of payment delays

Most invoice delays are due to organizational issues rather than a refusal to pay. Common causes include missing purchase order numbers, poor supplier referencing in the buyer’s enterprise resource planning (ERP) system, unvalidated receipt of goods by the warehouse, a validator on leave, a recent monthly payment campaign, or an unresolved disagreement on a specific invoice line.

Payment delays cost SMEs billions

A good practices guide published by the Observatory of Payment Delays in November 2024 classifies these organizational causes as recurring sources of delay. The Observatory estimates that French SMEs lose about 15 billion euros in cash flow annually due to payment delays, with a significant portion of this amount not due to bad faith but to operational friction.

Automated reminders may not be effective

Automated reminder sequences, while useful in reminding customers of forgotten payments, treat all invoices the same and can lead to customer annoyance, suspicion, or non-response. The work that truly matters—identifying the right contact, understanding a customer’s promise to pay by a certain date, and managing partial payments or disputes—requires human involvement.

AI improves communication and follow-up

Language models have changed the way businesses handle overdue invoices by being able to read and understand customer responses. A well-guided AI can adapt follow-up based on the customer’s message, ask the right questions in their language, and maintain a professional tone. It can also track the conversation over time, something overburdened teams often struggle with.

AI needs a strong foundation to be effective

The foundation of using AI in this context requires knowing which invoices are genuinely overdue after accounting for credits and partial payments. It also involves avoiding chasing customers in legal proceedings, ensuring no outdated information is sent, and maintaining a clear record of decisions made by people.

Human judgment remains crucial in key decisions

AI should not be used to make decisions that affect customer relationships and cash flow, such as granting a payment plan, adjusting a disputed invoice, or changing the tone with a strategic customer. These decisions require human judgment and should remain in the hands of people.

Before automating, ask critical questions

Before automating reminders, companies should ask three key questions: Do you know why your ten largest overdue invoices are not paid? Who decides when a customer proposes to pay in installments? What will remain for your team once the system is in place? Friendly recovery is first and foremost about maintaining a good relationship.

Timely payment comes from understanding, not chasing

Companies that get paid on time are not the ones that chase the most, but those that understand the earliest what is blocking payment. AI can help identify these issues on a large scale, provided it is used to understand before chasing and that people retain the responsibility for decisions requiring judgment.