What’s the deal? Deepentix, a deep-tech startup from Lower Austria, has raised a six-figure pre-seed round from tecnet equity, early-stage investor Rockstart, and business angel Kambis Kohansal Vajargah. The company builds an enterprise AI platform for regulated industries.
What’s the endgame? Deepentix aims to make every AI-generated statement traceable back to its source in the original document. Its platform turns documents into a knowledge graph — a verified network of facts and their relationships — so answers can be tracked to the exact sentence.
Why now? Companies in pharma, medical technology, insurance, and banking work daily with large volumes of studies, patents, standards, and internal reports. Standard generative AI delivers fast answers but often cannot fully prove where they came from. Deepentix targets that gap.
How it works. Rather than a single large cloud model, the platform uses several small, specialised language models trained on individual fields such as pharma or banking. Customers build, run, and adapt their own knowledge graphs.
“When employees have to blindly trust an AI because they cannot verify its answers, that is a real risk, especially in sensitive areas,” said co-founder and chief executive officer Morteza Amini (translated from German). “We are not consultants who build a knowledge network for individual clients, but a technology provider that gives companies full control over their own data and models.”
What’s the money for? The fresh capital goes mainly into research and development and into first pilot projects in regulated industries.
Peter Buchinger, investment associate at tecnet equity, said the founding team was decisive: “The combination of a scientific background and experience in international companies was the deciding factor in investing at this early stage” (translated from German).
The signal: As enterprises push generative AI into compliance-heavy sectors, verifiability is becoming a selling point in its own right. Deepentix is betting that traceability — not raw model power — is what regulated buyers will pay for.
Read more: Trending Topics
Image credit: Generated with Gemini