Martin Svane
Every payment has something to say. Most companies aren't listening.
A late payment is a signal. A support call after an invoice is a signal. A customer switching payment method is a signal. Every recurring payment carries information about friction, confusion, intent, even loyalty. But most companies have built systems that are deaf. The invoice goes out. The money comes in. Or it doesn't. Either way, nobody pays attention to what happens in between.
That in-between is where Billogram started.
It began with something simple: clarity. The invoice was redesigned not as a document, but as an experience. Who sent it. What the amount is. When it's due. Sounds obvious. But open your last utility bill and count the seconds before you find the payment deadline.
Then it became interactive. The digital invoice turned into a page where customers could choose how to pay, update their details, see their history. The payment went from something you process to somewhere you go.
But even the best-designed experience can't answer every question. And for industries like energy, telecom and insurance, where the product is invisible until something goes wrong, the payment moment is often the only touchpoint. Questions don't disappear because the design is good. They just change shape.
So a channel was opened. Customers could write directly on the invoice. Their message landed in the support team's inbox. Simple, almost naive. But it proved something important: people want to talk where the transaction happens. Not in a separate app. Not through a phone queue. Right there.
A chatbot followed. Then a better one. Machine learning handled the common questions around the clock. Payment status. Due dates. How to switch to direct debit. Useful, but limited. The bot could answer. It couldn't understand.
Today, Agentic Assist can. Built on a modern language model, it reads context, handles any language, and does something none of the previous solutions could: it acts. A customer asks for more time to pay. It grants a deadline extension. Someone questions a higher amount. It pulls up the previous payment and explains the difference. It doesn't just resolve the question. It prevents the next one.
And the payment still has more to say.
Because here's what fifteen years of listening has taught us: the conversation doesn't want to live on an invoice forever. It wants to meet the customer where they already are. In a message thread. A notification. A WhatsApp chat. The next step isn't a smarter invoice. It's a recurring payment that reaches out before the question forms. A reminder on the new due date. A nudge to activate direct debit. A heads-up when the amount will change.
Every payment has something to say. For fifteen years, we've been learning to listen. Now, we're teaching it to speak.
How does it actually work?
Read what the payment does when it stops stalling
Martin Svane