Successful Machine-Learning-Startup Targets Global Market


With MINT.extract, a solution to extract unstructured information from documents, turicode already generates more than one million Swiss francs in turnover. Until now, the team has relied on building the company via bootstrapping. For the next growth step, the team now wants to bring investors on board.


When asked what made him proud in his entrepreneurial life, turicode’s CEO Martin Keller only has to think briefly. He then states: "Up until now no turicode customer has ever cancelled a contract.” This serves as the best proof for the quality of the product according to the co-founder. In his view, the whole team must work together successfully to make sure that the customers get what they expect.

The start-up already won numerous well-known customers from Switzerland and neighbouring countries. In the beginning, the founders financed the growth by following a bootstrapping strategy to reach a milestone of one million Swiss francs in sales. "We put every cent we earned into the team," explains the CEO. Turicode has 16 employees, mostly software engineers and mathematicians.

Empowering end users to build machine learning systems

The team has developed the product MINT.extract, a machine learning based engine enabling companies to find and extract data from unstructured documents. 80% of all business-relevant data is available only in an unstructured format such as scans, pdfs or emails. “Until today you had to be a data scientist or at least be able to code to capture the potential of new technologies such as machine learning or deep learning. With MINT.extract, we empower end users to build their own machine learning systems with just a few clicks, for document and in any language”, Martin Keller is convinced. This opens up new possibilities in automating business processes and in enabling analytics.


In addition, the turicode product can be used for all document types. Many providers are limited to  specific documents such as invoices. But in Martin Keller’s experience, customers want a platform able to deal with documents of all kinds, since business processes always involve several document types. This is exactly what MINT.extract does. In addition, end users have the flexibility to train and extend the system on a continuous basis with just a few sample documents. This requires the team from Winterthur to be on top of research, which they demonstrated recently by outperforming international benchmarks for detecting and recognizing tables within documents.

There are years of development work behind the product. Three of the four co-founders are computer scientists who studied together at the ETH Zurich. Martin Keller, on the other hand, has worked in the financial service industry and as a management consultant for many years. So it is not surprising that, in addition to invest into tech and product development, the start-up puts a lot of emphasis on building a sustainable company with clear processes and structures in place "Recently we have passed a comprehensive risk assessment of a Swiss bank," the co-founder reveals.

Scaling up

So far, the sales team has brought the product directly to the market. In a next step, the company plans to implement a scalable sales and service model, including selling MINT.extract via partners and digital marketplaces in order to create an international distribution network. The high standardization of MINT.extract allows for a faster commercialization and strengthens the customer success team.

Turicode is currently seeking two million Swiss francs to finance the coming growth phase. Martin Keller considers it important that the investors do not just give money for a decent financial return but share turicode’s vision for the future. He says: "To drive growth forward, we are looking for investors who bring industry expertise, can open doors for us and are willing to contribute to our mission of turning documents into value".

Startup Pitches 

This is an article of our start-up pitch series. Until the summer break startupticker in collaboration with Swisspreneur will introduce Swiss start-ups on the search for investors.

(Stefan Kyora)

Picture - The turicode founders: Martin Keller (CEO), Benjamin von Deschwanden (COO), Aaron Richiger (Head Machine Learning),Patrick Emmisberger (CTO), from left to right.