System for automated distribution of support tickets to engineers assuring maximum customer satisfaction
Publication Date: 2017-Jun-06
The IP.com Prior Art Database
TITLE: System for automated distribution of support tickets to engineers assuring maximum customer satisfaction
Work of support teams serving many customers is governed by complicated rules where many factors
count. Typically support manager needs to deal with support skill gaps, customer skills, escalations,
minimize tickets' backlog, make sure engineers are not overloaded, assure there are no delays, etc., but
at the end only one goal needs to be achieved - customer satisfaction.
One of the more important activities is correct assignment of new support tickets to engineers in a way
which allows to maximize the customers' satisfaction. Current article provides a solution which can do
this in almost fully automated way.
We use neural network which learns on past support tickets to assigns best support engineer to work on
new tickets in such a way that customers' satisfaction is maximized. It also allows support manager to
steer/adjust the system parameters to make sure hot situations are properly covered and support
engineers are not overloaded. The system also allows support manager for manual assignment in cases
which can not be automatically covered be the tool providing all possible supporting information based
on which he/she can make the best decision.
The vectors being inputs for the analysis would consist of the following fields:
1. Customer contact email address
2. Company name (unique customer identifier)
3. Company current sentiment (based on sentiment analysis of all currently opened tickets)
4. Product area to which ticket is opened (code component) or alternatively ticket title
5. Ticket severity
6. Ticket priority
7. Escalation. For example, this can mean if critical situation is assigned. (true/false)
8. Number of tickets currently in backlog (this allows for correlation between backlog size and
9. Support engineer name (or unique identifier)
10. Number of tickets currently owned by the engineer
Please note that points 3, 8 and 10 are not trivial. Thanks to them we can eliminate time factor from the
considerations and focus on satisfaction only. Also analyzing current customer satisfaction from all
opened tickets allows to take into considerations and hot situations and escalations. Please note that all
this can be determined from archive of all past support tickets.
The analysis based on the input vector needs to determine the predicted custo...