Based in Ghent Tekst.com has developed a promising AI platform that enables large companies to automate text-based processes, such as processing high volumes of emails. Today, the company announced that it is raising €700,000 from Entourage to further develop and commercialize its technology across Europe.
Before the advent of advanced AI, turning text into usable data was something humans had to do manually. Consider processing specific information from an email to plug into a CRM, accounting, or ERP package.
The promise of AI is great, but generic LLMs (Large Language Models) like ChatGPT do nothing more than convert text to other texts. Accurately interpreting unstructured text and then converting it into usable data—an input that a database can work with—remained an impossible task.
Tiebe Parmentier, CTO and co-founder of Tekst.com said: “Existing software based on recognizing keywords and setting rules can automatically label about 20% of incoming mail traffic at an average company, with low accuracy. That is too little for companies to efficiently automate handling.” adding: “The custom-trained AI models we offer to our customers can figure out the meaning of many more messages, with higher accuracy. The fact that this eliminates repetitive tasks is not only excellent news for the companies themselves, but it is also positive for employees. We take ‘boring’ operations out of their hands, making their work more enjoyable.”
Major customers
The company has already won several major customers who want to get started with AI concretely and quickly. This includes US global medical equipment player Becton Dickinson and the large dairy cooperative Milcobel.
“Our enterprise customers have often already automated a lot of data processes,” says Tekst.com CEO Wouter Janssen. “But in their pursuit of ‘touchless automation’, they still face this last wall: processing large volumes of text in emails, cases, orders, and tickets automatically—but in a strictly controlled and secure way.”
Customer communication in B2B requires a careful approach, especially with large contracts, where fast and accurate communication is essential. Moreover, all agreements and rules on security and privacy must be taken into account.
Tekst.com’s AI model succeeds in mapping, specifically within the company, all the knowledge needed for the correct analysis and routing of messages. It effortlessly retrieves that information from existing data systems such as SAP and Salesforce. Furthermore, the work performed by Tekst.com happens in the background, so employees do not have to work with yet another new tool. Finally, Tekst.com meets the required standards for security and confidentiality, such as GDPR legislation and the global ISO 27001 standard of information security.
Christophe Boucquet, Customer Service Digitalization Manager at Becton Dickinson. said: “We have a large number of client-facing inboxes, and needed a solution that could quickly analyze and prioritize our growing email volume. Given the nature of our products, in some cases, our customers require extremely fast response times. With Tekst.com, we achieved results in just three weeks. With Outlook, it was impossible, and trying to train an AI model to work with email ourselves would have taken us many months.”
Pieterjan Bouten ex-Co-Founder and Executive Chairman of Showpad: “Tiebe and Wouter are hugely ambitious and driven by their goal of making AI work in complex business environments. They have developed a wonderful technology that delivers many efficiency gains in a very short time. We see huge potential for the company, and we believe they can grow into a global player.”
