AI chatbot for the automated capture of daily prices

In the international hazelnut wholesale trade, the pace is set not by quarterly plans but by daily prices. Anyone who wants to keep up has to check regularly with suppliers and producers and keep readjusting their own position. At Birlin Mühle, these conversations have traditionally taken place by phone and email, complemented by WhatsApp, which has become just as natural a channel in recent years.

However varied these channels are in day-to-day business, they become all the more demanding behind the scenes. Every price quote has to be followed up and filed properly before it is really usable for internal decisions, whether it comes in as a quick call, as an email with a price list attached or as a WhatsApp message. What looks like an uncomplicated dialogue with suppliers from the outside is, internally, a quiet but steadily growing administrative burden.

In day-to-day business, prices arrived in every conceivable form: mentioned on the phone between two appointments, as a short WhatsApp message, as a photo of a handwritten note, as an email attachment with the latest price list. Anyone who wanted to keep track had to work through all these channels, sort the information, transfer it into spreadsheets and, at the same time, remember which details were still missing or already out of date.

This revealed a paradox familiar to many medium-sized businesses: thanks to modern channels, Birlin Mühle had become much faster at obtaining information, but lost a great deal of time internally on following up and filing. An important offer could get lost among hundreds of messages, responses were delayed by hours, and in the end decisions were sometimes based on figures that were no longer quite current. In a market where daily prices are negotiated by the hour, the original speed advantage turned into an increasingly costly bottleneck.

For the management, the key question had long been clear: how can decisions be made faster and more efficiently without the team behind them continuing to work as a manual translation office? Together with Werkdigital, the decision was made not just to smooth out the process, but to rethink it from the ground up.

The answer was not to be additional software that forces a new interface on the partners, nor a superficial integration that only scratches at the symptoms. The plan was a concept that fully retains the familiar communication channels and quietly takes over structured data capture in the background. For the partners, nothing noticeable was to change in the process; internally, all the more.

Today, an AI-supported chatbot takes on the task that previously kept several employees busy at the same time. It independently requests the required daily prices from suppliers via WhatsApp and conducts the dialogue in a tone that suits the usual way of doing business at Birlin Mühle.

The responses arrive in exactly the form in which suppliers reply anyway: as a short text message, as a photo of a price list or as an attached file. The bot interprets them all, checks the values for plausibility and transfers them in structured form to the company's central database. If a detail is missing or a value looks unusual, it follows up politely instead of passing on incorrect information unchecked. For the suppliers, the familiar conversation experience is preserved, while a clean data basis that can be evaluated at any time is created in the background.

What used to be scattered across phone calls, emails and chat histories is now available to the Birlin Mühle team at the touch of a button. Market movements can be assessed faster and offers answered more quickly. Decisions once again rest on a data basis that no longer has to be laboriously pieced together.

In day-to-day business, this means noticeably less manual following up and filing. Information relevant to decisions is immediately available, and because the bot automatically checks details for plausibility, data quality has also risen markedly. The processes scale with the business without the administrative effort growing with it, and the team regains a great deal of time for customers and core business.

The project shows where AI delivers its greatest benefit in medium-sized businesses: not in spectacular visions of the future, but in automating the routine tasks that cost time every day. For Birlin Mühle, this has resulted in a reliable ‘AI employee’ that takes on a substantial part of the daily price work and relieves the team exactly where it used to cost the most hours.