Smart Ticket Prioritization
IT services support systems work on huge numbers of service tickets every day. While aiming to resolve most of the tickets on time, they are still facing heavy penalties due to SLA breach. For a major IT service support provider this loss usually ranges to 15-20% of overall revenue. The probability that the ticket will breach the SLA or get resolved on time is based on multiple factors, and very few skilled ticket dispatchers can manage to keep SLA breaches low. It is very important to identify the probability of service tickets that will breach the SLA as soon as tickets arrives as it helps to prioritize or assign ticket to proper resource.
Persistent has developed a machine learning based solution which considers all structured and unstructured data available within tickets as soon as they arrive into the service system environment, and predicts the likelihood of a new ticket to breach SLA.
Customer Success Snapshot
Customer is one of the largest IT service providing network, providing support services for their hundreds of products.
- To build a completely independent system to identify at-risk incident tickets accurately.
- This system should allow the dispatcher of IT support system to prioritize and use defined rules to reduce the risk of individual incident tickets missing important data.
- Categorize incident tickets into two groups: High Risk and Low Risk using classification methods which should leverage the unstructured text description included in tickets.
- A complete system with machine learning classifier to predict the probability of tickets to breach SLAs.
- This machine learning system used text mining process to identify the issues and problems faced by end user.
- The system included an accurate classification model based on the historical information of tickets resolved in time, based on tickets textual data and other structured data.
- The system was capable of fetching real time data and score classification models and showed the results on its UI.
- Reduced the overall penalty loss by 5-8%.
- 75-80 % accurate estimation of incident tickets as soon as they arrived.
- Reduced the heavy penalty cost on tickets breaching SLAs.
- Provided more nuanced work experience for ticket managers.
You May Be Interested In
Whitepaper: Facial Analysis: Face Recognition, Expression Recognition, and Gender Identification