SherpAI builds a learning data quality layer for enterprise systems. The software integrates into ERP and CRM systems, identifies faulty data structures, corrects them in context and completes missing information – an area where research artefacts developed at this chair move directly into practice.

From Research Artefact to Product

Entity matching and data quality in enterprise systems have been a research topic here for years: how can an AI system decide reliably whether two records describe the same real-world object, and how can that decision be made comprehensible? SherpAI turns this line of work into a product by combining several AI paradigms in a modular, orchestrable architecture instead of a rigid rule engine.

The SherpAI platform: analysis, cleaning and enrichment

The Role of the Chair

Prof. Dr.-Ing. Marcus Grum acts as originator of the idea and as mentor to the founding team; the venture is being built out of the University of Potsdam. The twelve-month EXIST phase is dedicated to moving from demonstrator to a market-ready product, with ERP manufacturers as integration and distribution partners.

For the chair, the spin-off closes a circle that our transfer strategy has aimed at from the start: research results should not end in a paper, but in systems that people actually use.

More about SherpAI