Tasks of automatic incident handling
Support services of large companies deal with a huge flow of requests and incidents. To radically improve their work, you need to automate some tasks:
- Detect mass outages and service failures in real time. Inform the dispatcher and support manager about them, so that the necessary measures can be taken in a timely manner.
- Reduce incident routing time. Automatically determine when an incident is logged whether it is a typical, mass or simply a known issue.
- Offer a ready-made solution if the incident relates to a known problem.
- Obtain up-to-date information to analyze the quality of supported systems and services. Regularly generate reports to manage product quality and properly distribute the efforts of developers and support staff.
To effectively automate incident handling, two interrelated technical challenges must be addressed:
- Clustering. Identify a list of incident clusters, including groups not known in advance that may result from outages and mass service problems.
- Classification. Check whether each incident belongs to one or another group.
NLU technologies for incident handling
Such tasks are usually solved by machine learning (ML) tools. However, it is not so easy to apply ML methods for processing calls to technical support. Different words may be used in descriptions of incidents that are close in meaning, and different incidents may have the same words.
To categorize incidents in a meaningful way, it is necessary to extract the essence from their textual description. NLU (Natural Language Understanding) technologies help to achieve this. Processing incident descriptions with NLU tools involves several steps:
- analyze the grammatical structure of sentences
- Identify the lexical basis of each word
- search for concepts - formally defined terms
- making a “semantic portrait” of each phrase - extracting formal statements.
How does EKG Language Processing work?
Such an algorithm will evaluate the commonality of incidents, even if they are described in different words. It “understands” negation and modals such as “may”, “should”, “requires”.
We introduce DataVera EKG Language Processing, a natural language processing tool that is part of our platform. It can be used for clustering and classification of incidents. The tool integrates with any Service Desk system and is deployed in the customer's infrastructure. You don't need to specify groups of incidents and typical problems in advance - the service will define them itself. You can manage the work of the service by “helping” it to identify incidents that are similar in meaning, or by “showing the difference” between incidents that are similar but different from the business point of view.
Processed data in the platform interface. The incidents that have been grouped are linked to typical problems. The names of typical problems are generated automatically using summarization. In the properties of each incident we can see elements of structured descriptions of its meaning, which are used when comparing it with other incidents. These descriptions contain references to the concepts mentioned in the incident description. The platform interface allows analysts to explore the results of incident classification and manage the work of the algorithm. The work of technical support specialists takes place in the familiar Service Desk system, where the classification results are exported.
If it's important for you to improve your helpdesk performance and problem management, contact us. We'll give you a demonstration and discuss how we can help you solve your problems.
