Free text answers are not only a common and popular means of survey research, they are also a frequently used instrument of many touchpoints in companies (hotline, ratings, general feedback, CRM data etc.). Such open responses provide you with important insights for controlling relevant KPIs, as comments, opinions, experiences, etc. are given in your own words. However, traditional approaches to the analysis of such text fragments are very expensive and complex and pure machine-algorithm controlled tools often do not deliver the results in the desired quality. For time-critical processes, such as the automated classification of critical incidents, efficient and reliable algorithms take over the entire process.

LINK uses a combination of both approaches for processing text data: Modern algorithms and efficient processes combined with professional multilingual human coders. The basis for this is over 30 years of manual experience with code frames combined with artificial intelligence for machine learning. This novel form of human-machine interaction leads to a cost-effective way of processing large amounts of text into the desired codes: The human trains the machine, the machine finalises the work. You can decide for yourself how large the proportion of human coding in the process should be. Once the machine has been trained, the algorithm is set and newly arriving free text responses can be automatically encoded again and again.