Textgain was founded in 2015 and had as its core product a text analytics API. They have been growing over the last years and now also do custom NLP classifiers for customer-specific purposes.
Guy De Pauw, CEO & Co-Founder of Textgain, was kind enough to answer some of our questions and tell us a bit about how the company was born, how they set apart from others with similar solutions and how Textgain has a team of cat lovers, that also welcomes dog lovers 😉
What’s the story behind Textgain? How was the idea for this business born?
We founded Textgain in 2015 as a way to bring the exciting technology that we had developed at the University of Antwerp to the market. The good thing about academia is that you get the freedom to work on innovative technologies, but at the end of a research project, the technology you’ve developed is usually far from production-ready. At the same time you don’t want to see it disappear into a drawer, either. That’s why Tom, Walter and myself founded Textgain: we wanted to actually see our innovations being used in practical applications. And what better way than to just do it yourself…
Upon Textgain’s launch, our core product was a text analytics API that could perform state-of-the-art Natural Language Processing tasks for a wide range of languages. Stuff like concept extraction, sentiment analysis, … But our USP was the author profiling technology that could automatically extract demographic information about the author of a document, based on his/her writing style. This was actually pretty far out there in terms of what you could do with NLP, and we started to notice that people were struggling a bit to understand exactly what our NLP-pipelines could do for them.
With Gijs and Redouan, we added two business developers to the founding team and started to pivot our activities to improve our product-market fit. The core product was still the text analytics API, but we also started to use the underlying pipelines to develop custom NLP classifiers for customer-specific purposes. And that’s when things started taking off, especially when we started to develop detection technologies for hate speech and radicalization that got a whole lot of traction in a short period of time.
What sets you apart from others with similar solutions?
AI has become a very hot topic in recent years and unfortunately this has attracted a lot of malpractice to the field. This seems particularly the case for the AI-subfield of Natural Language Processing, with many companies overpromising results and charging their customers a fortune for the privilege.
At Textgain, we believe that the NLP development and deployment process should be honest and transparent. True to our academic roots, we ensure that our customers are made aware of the possibilities, but more importantly also the limitations of the technologies that we can offer and develop for them. While this may not seem as the best strategy from a commercial perspective, we believe we have a role to play as NLP evangelists. This no-nonsense attitude has earned us the trust of high-profile clients, many of whom have had prior experience with overpromising vendors and appreciate our scientific approach to tackling NLP problems and the robust results that we deliver.
I’m sure Textgain is more than a deeptech company focused on the technical aspects. Tell us a bit about the people behind it and what are the most important human aspects you search on someone when considering them to be a part of your team.
As a commercial NLP company it’s important for us to have a differentiator over our competitors, we are not so much looking for computer scientists that know a little bit about linguistics: we are looking for linguists that know how to work with language in the context of Artificial Intelligence. Our current data science team fits that profile perfectly. Just as important are the people who know how to apply our technology and we’re very blessed to have people on the team who are interfacing with civil society and law enforcement so that the tech team is not developing stuff in a vacuum. It’s great to work with people from widely different backgrounds.
One common denominator seems to be our collective love of cats. Last time we counted there were 12 Textgain cats. This is also reflected in our artwork for the DTCT project and our software products oncilla and ocelot. But I have to add that we will also welcome dog people on the team 😉
And about MME… What were the greatest learnings you took from this programme?
We really enjoyed interfacing with our mentor Roderick Lindner, who gave us a massive amount of tips and tricks on how to establish and sustain growth. The sessions with Roderick were very inspirational and I think they will resonate for years to come. The MME-workshops were also great, particularly the selection of topics. Even those that were not that relevant to us at this point in our life cycle, were still useful as a refresher. And we especially enjoyed the interactions with the VRT-sandbox team who kept us motivated and inspired throughout.
Textgain is one of the startups that integrated the first support cycle of the MME programme. Find out more about them on www.textgain.com/.