Endress+Hauser after-sales service works with a CRM system, where a large amount of knowledge articles are available to resolve customer enquiries. In order to ensure that high quality standards are met, knowledge articles are periodically reviewed and evaluated by KCS Coaches based on the Article Quality Index (AQI), as proposed by the KCS guidelines. However, in the current situation, only excepts of all produced knowledge articles are looked at in a randomized fashion, since this is a labour intensive task and no partially automated tool is available.

The status-quo of the AQI process within Endress+Hauser was analyzed. Every AQI was looked at to assess whether automation is technically feasible using Natural Language Processing (NLP) methods. A selection of four AQI (Content Clear, Avoid Internal References, Uniqueness and Title Reflects Article) was then automated. The results were output as a human-readable report and used by a KCS Coach from Endress+Hauser. Feedback from the KCS Coach was used to iteratively optimize the algorithm.

A partially automated AQI tool was developed which increases the efficiency of the KCS coach. The tool helps to reduce the time needed to evaluate each knowledge article. Thus, a higher amount of checked knowledge articles will be available inside the CRM knowledge base for the customers as well as for internal consumers of the knowledge. The tool is also be used to analyze KPIs to compare the performance of entire Endress+Hauser entities, which enables the more targeted allocation of employee development and training initiatives.