go to top
Obsidian Business Service Intelligence

Automatic interpretation of service management indicators

Presentation at GigaTIC2017 on artificial intelligence techniques for the automatic interpretation of service management indicators

Presentation at GigaTIC2017, event organized by iTSMF and ISACA, of the research on the automatic interpretation of service management indicators implemented in Obsidian. Realized by the professor of artificial intelligence Alberto Bugarín and by the director of Obsidian, Diego Berea.

The video currently has 37067 visits on YouTube and 358 likes.

If you want to stay informed of new videos posted by Obsidian, subscribe to the channel, and YouTube will periodically send you notifications when new videos are posted. The Obsidian YouTube channel currently has 2768 subscribers. 

Data2text implementation on Obsidian

Within the disciplines of artificial intelligence, there are several related to natural language processing (NLP). Of these, a subset are the techniques of automatic generation of natural language (NLG).

Since some years, it had been done an intense research work in these areas. As a result, there have been researchers developing and deploying different software. These software have been tested in real environments, with satisfactory results.

Especially relevant is that considering predictions for 2018, NLG is going to be present on a large percentage in the BI market tools.

In this scenario, the management of ITSM services offers a very rich context for the application of the NLG techniques. It has multiple actors (technicians, managers or directors). Each one with different information needs. These can be technical specifications, service performance values or simply know if the service level agreed with the customer have been accomplished or not. That information can be extracted from the corporate tools of service desk and monitoring. NLG techniques attempt to help user, by processing data and displaying a written interpretation of these data according to user’s role.

Initially, this presentation exposes the most common methodologies to implement an NLG system of type data2text. At the end, are shown multiple examples of practical application to the automatic interpretation of service management indicators. The examples are based on the data2text implementation performed in Obsidian. This implementation includes both service desk metrics and availability, capacity or service level indicators, which is already working in real environments. With data2text implementation is now possible to generate written automatic reports using data obtained from these tools. This is one of the main advantages of these types of systems: reduce analytical effort that have to do the user.

More information on the data2text functionalities of Obsidian.