Skip to main content

Classification Of IT Helpdesk Tickets

Client:  Large IT services Company

Business Problem:  Internal IT Infrastructure, Applications, Network and Security support staff is high and also SLA scores low.  Need solutions to decrease the  head count reduction and increase the SLA adherence.

Data Challenges:   Past history data of Tickets, classifications, comments from support person and finally ticket closure comments were very parse.  Lots of e-mail tickets were also sent and attended to as per ground practice.

AI Solution:

  1. Ticketing system data of the past three years was collected and a subset was manually reviewed and enhanced.
  2. Domain experts created synthetic data for all key classification of tickets.
  3. Additional synthetic data was generated by auto generation tools developed in-house based on the experts data.
  4. All these data were used as Training data to train the classification engine.
  5. The AI Engine was integrated with current Ticket management system hence requiring no switch over to another system by the L1 support staff.
  6. The incoming tickets were analyzed by the classification engine and for all tickets with high confidence level of classification prediction, automatically assigned to various L2 support teams.
  7. Easier tickets such as “Reset my Password” or “ Give access to certain system” were automatically executed by using pre-defined scripts and the ticket was closed without any human intervention.

Solution Benefits:

  1. More than 70% of the tickets were automatically classified.
  2. All the simple tickets fully automated from end-to-end.
  3. Saving about 40% of L1 support staff
  4. SLA times improved by removing a L1 queue wait time and L1 analysis and classify time, thus improving overall SLA performance.

Note:  These case studies are experiences and accomplishments by our Think-Tank members