Put an end to notes at meetings

A constantly changing world, automation in everyday life and yet we sit in meetings for hours every day and shortly afterwards no longer know what it was all about. The founders of Tucan.ai Florian Polak, Michael Schramm and Lukas Rintelen want to have found a good solution for this. In this interview, they talk about how they came up with their idea, why it is so unique, and why falling flat on your face is simply part of the game. Besides, we tested their software directly and were surprised.

Why did you found your start-up?

There are many other companies that want to make meetings better with very different approaches. Most of them basically offer a transcription of the meeting, where you then get a transcript after the meeting. The problem is that if we talk for half an hour, we end up with 8-10 A4 pages, depending on how fast we talk. In the end, no one ever looks at it again. We focused on reading out the most important information from the conversation, in addition to the transcript, and then presenting it clearly on one page. The whole thing works by means of speech recognition. In good German, this is called "Result Protocol". At the end of a meeting, we provide both the verbatim transcript and a summary of the most important information from the meeting. The summary can then be easily sent to all participants of the meeting and even people who missed it will know what the most important information from the meeting was.

How did you come up with the idea?

Originally, we did something a little different, and that was a podcast platform called RecTag. Here we already noticed that we had more and more meetings in one day. Basically, we were sitting in pure meetings for up to 18 hours a day. The bigger our team got, the more meetings we had. Corona made this even worse, as the inhibition to have online meetings decreased further. We then quickly started working internally on a way to improve this, and then last year decided to focus on this and push the technology in this area. We quickly got a lot of interest from customers and that's why we're now where we are today.

In a nutshell, how does Tucan work?

The whole thing works two-sided, meaning that our product can be used as a cloud solution on the one hand or (for larger projects) can also be installed in the customer's system. The latter is only possible because we have built the technology ourselves and in that case it is always personalized to the individual user. The participants of the meeting can control the recording. They can start and stop it but also add things like "Hey Tucan, summarize the quote". Tucan learns with through the feedback you can give it after a meeting. So with each meeting, Tucan becomes more customized to the user's needs.

You received funding last year. Could you maybe tell us how you managed to convince the investors?

It was a long and hard road. As an AI startup, you have a long preparation time and a lot of development work. That means before you can even start, you theoretically need money. In the beginning, we had APX as an investor, at that time still as an accelerator from Axel Springer and Porsche. With relatively little money, we then developed the program, which was quite difficult. We work on difficult AI problems (speech recognition on the one hand and summarizing texts on the other). Summarizing texts well means for the program to understand linguistic context, which is very easy for us humans, but incredibly difficult for a computer. The exciting thing about this topic is that we are on the cutting edge of computers actually being able to understand us humans. The implications of this are enormous! Over the course of the last few months, we were able to acquire our first customers, who in turn recommended us to other customers, and so our credibility grew very quickly. In June of this year, we closed our first investment round and attracted many investors, including well-known ones such as IBB Ventures, the Spanish VC investor Faraday and Telefónica's corporate VC Wayra, as our main investors.

What made your customers recommend you to others?

This was mainly due to two reasons: our product is much more accurate than other products on the market, especially in German (the conversations are correctly identified & the summary captures the most important points) and we are usually very fast at resolving customer feedback. Especially in the beginning, no software product is perfect. However, we have been able to resolve some issues particularly quickly for our customers (customer writes to us and the issue was resolved in under 15 minutes in some cases). Our learning here is that you can provide excellent service even with a less than perfect solution, especially in the beginning. The downside of this is having to work late evenings or weekends.

Could you describe one mistake you made in more detail that other startups could learn from?

One very important learning that we learned only after making a few mistakes is how to properly interview users. We used to ask testers a lot of hypothetical questions (How would you use us? What else would you need for features? etc.). The problem with hypothetical questions is that you never come up with valid learnings that way. Instead, today we always just ask about their current process to better figure out how our solution can add value here. A book I wish I had read much earlier on this topic is Mom Test.

Which tools do you use in your daily work besides Tucan that might help others?

Our core theme with Tucan is to automate strenuous and repetitive cognitive tasks. Another tool we use and are super happy with is Getmoss, which automates many accounting processes. By using it, we are saving ourselves the position of an in-house bookkeeper for the foreseeable future. Really great!

What are your next steps?

Currently we have 16 people on our team, but we want to expand it even further and spread to other countries. Currently we are looking for another person in business development to improve the acquisition structures.

Where do you see Tucan in five years?

Basically, we believe that in the next five to ten years the working world we live in today will fundamentally change. All repetitive, cognitive work steps will be automated. The majority of our working time will simply be spent on things that, in principle, a machine could do better. In other words, we believe that through automated processes such as Tucan, we should all theoretically work less. Through Tucan, we will do our part because we focus on meetings and the idea of covering the meeting spectrum. We want to make life easier so you can focus more on the things you enjoy and have fun with.

Thank you very much for the interview.

Our conclusion about Tucan:

After the interview with Lukas and Florian, we received the transcript of our conversation via Tucan. It reproduced almost everything spoken 1:1 and highlighted the most important points. Like every application in the testing phase, Tucan is not without bugs yet. Nevertheless, it offers an incredibly good basis to recapitulate meetings afterwards and to send a protocol with the most important facts to all participants without any effort.