More control and less work. Even in your industry.
Fast and secure invoice processing for every organisation.
Fast and secure invoice processing for every organisation
The possibilities of our platform
In this article, Sander Kamstra—Managing Director at Payt—shares how we’re actively integrating Artificial Intelligence (AI) into our credit management software. As the job market continues to tighten and business processes need to become faster and smarter, we believe AI will play a vital role. Not as a replacement for human interaction, but as a powerful support tool to help companies work more efficiently, get paid faster, and deliver better service.
Table of contents:
At Payt, we’ve been actively experimenting with Artificial Intelligence (AI) for several years. But let’s take a step back and look at our vision for the future of credit management—and the role AI will play in it.
Imagine an AI worker—we call him Payter—functioning in the near future as the smartest team member you could wish for. He communicates flawlessly, learns continuously, and has no limits to what he can be taught. Payter thinks faster than we do, never gets tired, and never calls in sick.
The latest generation of AI workers can even define their own goals and learn how to achieve them as efficiently as possible. That makes AI incredibly powerful—but also a bit risky. If AI makes a mistake, it will do so just as confidently as when it’s right. This phenomenon is called “hallucinating.” That’s why a strong feedback loop—where AI constantly learns and is adjusted—is essential to ensure reliable results.
At the same time, we’re all witnessing a growing shortage in the labour market. The future workforce for the next 25 years has already been born, meaning the number of working professionals will continue to shrink. That presents a real challenge. And let’s be honest—not many people enjoy chasing down unpaid invoices. This is exactly where AI provides a smart and timely solution.
In this article, I’ll walk you through how Payt is already using AI today to make credit management smarter, faster, and more personal.
If you want to know when an invoice is likely to be paid, AI is incredibly good at making that prediction. Processing large volumes of data and building reliable forecasting models is where AI truly shines. And the best part? The model keeps getting smarter—continuously learning and improving as new data comes in.
You can turn these insights into smart automations right away. For example, if the Credit Risk module predicts that a particular invoice has an 80% chance of being paid late, you can automatically trigger a friendly reminder before the due date. Or you can set reminders to be sent in quicker succession. You could even deploy an AI worker to proactively call the customer.
This way, you combine the power of data with intelligent workflows and stay in control of your cash flow—before issues even arise.
Below is a screenshot of how AI-based payment predictions are visualised in our application:
Payment prediction is currently available in Payt and can be requested through our service desk.
AI is now capable of making calls to your customers. Of course, it’s important that the customer is informed that they are speaking with an AI worker. The AI politely asks for permission to continue and explains that the call may be recorded for training purposes.
Take Payter, our AI worker, for example. Payter might call to ask whether an invoice from Company X has been received. Since many emails—especially to addresses like Hotmail—often end up in the spam folder, Payter can guide the customer through whitelisting the email address to prevent this from happening again.
Naturally, Payter will also ask if there’s anything else they can assist with. If there is, Payt automatically creates a task for a human colleague at Company X to follow up. Every call is summarised neatly, and the full audio recording is logged in the timeline.
We’re planning to expand the use of this technology across more stages of the process. And the best part? Payt customers can choose exactly where they want to deploy an AI worker within their workflow.
Every day, tens of thousands of messages flow into the inboxes of our customers. And that’s a good thing—because strong communication between suppliers and customers leads to faster payments. But all that interaction also brings a significant workload.
One important innovation that AI offers here is the Large Language Model (LLM). This technology doesn’t just understand the words you use; it understands the meaning behind them too—and it can communicate independently, neatly, and effectively. Sometimes even so neatly that it feels almost unnatural.
Interestingly, tests show that people often quickly build a connection with an AI model. Where the line between real and artificial interaction lies may be a philosophical question, but one thing is clear: AI makes communication smarter, faster, and more personal. Exactly the kind of progress we aim for at Payt.
Below you’ll find a screenshot showing how our AI functionality automatically labels incoming messages within our software:
With this feature, incoming messages are automatically labeled based on the sender’s intent. Here’s how it works:
This gives you better visibility across all received messages. For example, you can quickly select and archive all confirmation messages at once.
The message labeling feature is now available in Payt and can be requested through our service desk.
In my personal view, within five years, all successful companies will be using AI in customer support. The advantages are simply too great to ignore. Of course, some hesitation remains—people often find it difficult to fully trust a computer. Every mistake tends to reinforce that scepticism. Think about self-driving cars: they already perform better than human drivers on average, yet every incident immediately sparks a “See?!” reaction.
Fortunately, the consequences of an occasional error in support are far less dramatic.
Starting in 2026, Payt customers will also be supported by Payter, our own AI support assistant. Payter draws his knowledge from our existing support articles. If he doesn’t have an answer to a question, he automatically creates a task for our human support team to expand the knowledge base. Our support colleagues actively monitor Payter’s responses and provide feedback whenever needed. And of course, customers can always choose to switch directly to a human agent.
In the beginning, there will still be many questions that we consciously continue to handle manually—for example, actions like deleting groups of invoices. But over time, more tasks will be handed over to Payter.
Thanks to Payter, we’ll soon be able to offer support even outside regular working hours. And an added benefit: by using AI smartly, we can keep our team small, highly skilled, and ready to grow—even as the labour market tightens.
