Recruiting
Language software analyzes applicants
Using language to find the right applicant? Yes, say the developers of the 'Precire' software. Artificial intelligence and machine learning make it possible.
Every person is different in terms of their personality and the way they express themselves. Our choice of words and way of speaking influence the way we think and communicate at every moment of our lives and therefore all areas of our lives - including at work. Software could help recruiters. The 'Precire' software is a technology that uses artificial intelligence to identify patterns in spoken and written language and derives linguistic, psychological and communication-related characteristics from them. In addition to natural language processing methods, specific text patterns (word combinations, word sequences, sentence structures) are also recorded. According to the developers, these patterns are then used to train objective prediction models based on reference data sets. These predictions are processed in aggregated form and deliver high, measurable added value in various business disciplines (HR, CX, e-health), the Aachen-based company promises. Customers include: Fraport and Randstad.
No replacement, selection
Numerous validation studies (internal, external) safeguard the process. A scientific advisory board supports the company. This includes researchers from LMU Munich and RWTH Aachen University. The software groups together interviews of people with different characteristics, for example, all people who exhibit certain characteristics or communicative modes of action. It recognizes which speech patterns are typical for the respective groups. High-dimensional similarity comparisons are then carried out and predictions are made as to which of these groups a new, unknown person to be described is most similar to. The reference data sets consist of spoken language (interviews, calls) and written texts (e-mails, chats) and contain a large number of external criteria (KPIs) and additional information.










