link to top

Basic information for the media

About Cognotekt

Cognotekt specialises in semantic text analysis with the help of artificial intelligence.

Computer systems (“Turing machines”) offer an enormous potential to delegate repititive text-based administrative work to computers. In the course of the development of powerful language technologies (Natural Language Processing - NLP), this promise of automation is now coming within reach.

With its proprietary Wernicke® technology, Cognotekt has developed one of the most powerful mind-mapping natural language processing technologies available today. The extremely low error rate in categorization of less than one percent predestines Cognotekt’s technology for real unsupervised automation.

Opposed to the view of some critics of technical progress, we do not consider automation dangerous but rather see it as beneficial: employees can attend to more demanding tasks while the software performs simple routine tasks more cheaply, faster and with greater repeatability. Many applications are only possible through ai based text analytics, for example the search for compliance violations in large data storages.

We think that in this way administrative work will be more productive and more satisfying and the demographic gap on the labour market can be closed.

This idea is what drives us.

Dates & Facts

Employees (persons)ca. 35 (ø 2019)
Employees (full-time-equivalents)ca. 25 (ø 2019)
Legal statusGmbH

About our natural language processing technology Wernicke®

Automatising intellectual work requires a precise understanding of language. For this reason, we developed Wernicke®, our own speech technology, named after the neurologist Carl Wernicke who discovered the speech centre in the brain.

Wernicke® is a linguistically trained AI which highly accurately extracts information from free texts. Other than AIs that are based on neural networks or other stochastic models, Wernicke® does not only rely on statistical processes but determines meaning through exact logical inference. Cognotekt developed a process to transfer language into mathematical logic which is globally unique. This differentiates Cognotekt technologically from other companies in the world of AIs.

Besides the unsurpassed language comprehension, the Wernicke® technology has a second advantage for concrete appliances: Applications based on Wernicke® require significantly less training data than AI applications that are purely based on stochastic models. Cognotekt’s solutions such as Eloquent®, Argument® and Konsequent® are therefore also suitable for small and medium-sized companies that do not have huge data records at their disposal. Cognotekt thus provides a remarkable key to success which offers independence from technology offered by market leaders in the AI business.

Vita Jobst Landgrebe, M.D.

Jobst Landgrebe M.D., born in 1970 in Cologne, studied medicine and biochemistry in Heidelberg, Göttingen, Basel and Nîmes. Already in his adolescence, he programmed IT systems. In 1998, he finished his doctorate’s thesis and until 2002 he worked as a neural geneticist at Max Planck Institute for psychiatry in Munich. He studied mathematics part-time in Munich and Göttingen.

From 2002 to 2006 he worked as a research assistant with biomathemathics as the core research area at the centre for cell biology of the Deutsche Forschungsgemeinschaft at Göttingen university. Besides his research activity in mathematics, he dealt with theoretical computer sciences and turned away from medicine. From 2006 onward he worked as a business consultant at Booz Allen Hamilton and as a software architect for artificial intelligence at a provider for medical informatics Clinerion.

In 2011, he founded a first AI-enterprise for a medical software system as a entrepreneurial company. In 2003, he founded Cognotekt GmbH.

In 2019, he published the paper “Making AI Meaningful Again” as a joint work with the British philosopher Barry Smith (University at Buffalo, NY). The paper was peer reviewed and accepted by the journal for philosophy “Synthese”. The paper critically scrutinises the suitability of today’s AI systems (DNN, machine learning) for semantic language comprehension and offers an alternative model which is based on formal logic.

Jobst Landgrebe, M.D. is married and has five children. He lives with his family in Cologne.

Publication “Making AI Meaningful Again” download (in English)

Contact for the media

Christoph Lieck
Head of marketing & communication
+49 221.643 065 10