There are moments when a chance discovery changes the course of events. For me, Kevin, it was a scientific paper from Australia on Varroa detection through frequency analysis. As a budding business IT specialist, I was immediately fascinated by the idea of using machine learning to detect problems in a beehive. What began as a university project became HiveSound—albeit in a slightly different way than we originally thought.
The basic idea was clear: to develop a monitoring system for honeybees that would help Beekeepers identify problems in their colonies at an early stage so that they could respond in good time. Together with my former Roommate Julian (software developer) and Michelle (biologist), an old friend of Julian's, we entered the AI.STARTUP.HUB Ideation Program with this rough vision.
The team agreed that the idea had potential—and we had the various skills needed for such a project: AI expertise, software development, and biological know-how. While I was still fine-tuning the last details of my master's thesis, we were already preparing together for the next step to start working on the project full-time: applying for the EXIST scholarship, which we received a short time later.
During EXIST funding, we did what you do as a founder: conduct interviews, build prototypes, test, question. We spoke to countless apiaries and developed our own sensor prototypes and came across a very practical problem.
In order to label and train our sensor data in a meaningful way, we had to correlate it with actual events in the beehive. That meant we needed a way to make it as easy as possible for beekeepers to document their reviews. So we developed the first software components — initially only as tools for our actual goal, the monitoring system.
Julian and I decided to start beekeeping ourselves. After all, we wanted to understand what we were actually building for. And this is exactly where something decisive happened: We became painfully aware of how impractical all existing solutions are.
Imagine: You're standing at the beehive, wearing thick gloves, the bees buzzing around you, you have to work fast — and then you have to type in some data into your smartphone? It just doesn't work. The reality of beekeeping and the available digital tools were miles apart.
At our first Eurobee trade fair in 2024 as an exhibitor, we also evaluated these software components — the focus was still on the monitoring system. But the conversations with beekeepers showed us something important: We were not alone in our problem. Our surveys and interviews confirmed that many beekeepers experienced exactly the same frustration.
It was one of those decisions that feels logical in retrospect but is difficult right now. After Eurobee and as part of our preparation for the application to InnoFounder, a program by Innovations- und Förderbank Hamburg, we have decided to focus on the software solution first and take the issue of monitoring a step backwards. A small pivot in the sense of order — but one that felt right.
Why Because we have seen that we can solve a real, acute problem that affects almost every beekeeper at every inspection and aftermath. The best hardware solution is of no use if the basis — practical documentation — doesn't work.
Almost a year later, after we were successful InnoFounder After months of development and intensive exchange with beekeepers, we brought the software to app stores and the web.
The core: A really powerful voice assistant that automatically fills out the entire stock card from spoken text, creates todos and creates notes. This means that observations can finally be easily recorded while standing on a hive — even with gloves, when windy, or a swarm of bees around you.
And the issue of monitoring? We have not forgotten this, but have now gained strong partners whose systems — stock scales (e.g. HoneyLink), weather stations, sensors for the brood nest — we can easily connect via our software and IT infrastructure. The sensor measurements are thus presented centrally in the HIVESOUND app and stored for access in addition to your own reviews. With our closest partner HIIVE Link, we have been working together for several months now to bring an innovative monitoring system to market at a fair price. More about that soon!
Our relatively fresh AI assistant (currently in beta) goes one step further. He can react to all stored data via the in-app chat function — whether it's reviews, sensor data or notes. Thanks to our manually maintained beekeeping knowledge database, he can answer general questions, but also to very individual inquiries about your own apiary by connecting to his own transparent/sensor data.
Imagine asking, “Who are the weakest people in my locations right now?” or “Were there any abnormalities at the location during the last inspection meadow?” — and the assistant analyses your reviews to give you a well-founded answer and provide you with the best possible support.
In the coming months, we will go one step further and adapt the assistant to more complex issues. More specifically, this means an analysis process in which recorded data and external sources (such as weather reports) are processed in detail and combined with each other in order to be able to provide analytically sound conclusions on input questions.
Where do we want to go? We have a clear vision: HIVESOUND should become a standard in beekeeping. But not as a closed system, but one with open interfaces.
In concrete terms, this means:
For beekeepers: Complete freedom in choosing sensors — buy where you want, or even build it yourself. No vendor lock-ins, no forced hardware purchases.
For research: A unique database of manually managed reviews and sensor data, which we — after user consent — can share with research institutions, schools, universities and joint projects.
For the future of beekeeping: Digital beehive twins that identify problems before they arise. Identify the swarm atmosphere before half the people move out. Be able to improve winter survival rates and accurately forecast honey yields.
When we look back at the journey from the Australian research paper to today, there are three lessons that stand out:
1. It's the market that decides, not your original idea. We started with hardware and ended up with software — because the problem was more acute there.
2. Become a user yourself. It wasn't until we became beekeepers ourselves that we understood the real problem.
3. Openness wins. Instead of building our own silos, we rely on partnerships and open interfaces. Because in our opinion, we only get something moving in the apiary if we stop cooking our own soups.
We're still a long way off. The beta is running, the first beekeepers use HIVESOUND every day, and we learn with every feedback. The vision of digital twins is still a long way off — but it is getting closer with every day that a beekeeper documents his review via voice command and a sensor delivers data.
Honey beekeeping needs our help — not only because of the Varroa mite, but also because of climate change, pesticides and habitat loss. If we can make a small contribution with HIVESOUND to make their work easier for beekeepers and at the same time create better data for research and biodiversity, then every rocky moment of this founding journey was worth it.
Interested in how HIVESOUND can support your apiary? Visit us on hivesound.ai, follow us on instagram or write to us by email. We are looking forward to the exchange — whether as a beekeeper, hardware partner or simply as someone who is interested in the future of beekeeping.
About the authors: Kevin is a business informatics specialist specialized in AI systems, Julian is a senior software developer and IT specialist. Together, they founded HIVESOUND GmbH — and learned how to beekeepers along the way.
@HIVESOUND 2025