Ticket Next Best Action
AI for issue resolution
MetTel, like many large companies, has a trouble ticket system in place to respond to customer needs. Each ticket is sorted and resolved by an operator or engineer. Due to human variance, issue resolution can vary. MetTel wanted to ensure consistency in quality and speed when resolving tickets, and was looking for a way to improve profitable growth.
What we did
One of the key features of any ticket management system is the “Next Best Action,” an approach that requires operators to select the most appropriate action each time they manage a ticket. Using MetTel’s historical data, we trained an AI to learn best practice when it came to selecting the next best action.
The Ticket Next Best Action (or TNBA) AI now replaces the human decisions involved in determining the Next Best Action. This breaks away from a traditional, linear decision tree-based model, making MetTel’s solutions efficient, yet flexible and natural. It ensures every ticket is treated with the same quality as MetTel’s best engineer, every single time. Consistency of quality is a key element to TNBA. We also automated the system’s response, helping MetTel to achieve zero human touch ticket resolution.
We trained an AI with the vast amount of data MetTel already had at hand, using Natural Language Processing to extract ticket context and subject. Then, our predictive models were applied to a resulting structured data set along with hundreds of additional ticket features.
MetTel can now handle a mountain of tickets without growing proportionately in staff size – meaning their people and costs are now detached from their growth. They deliver more with less using existing resources to ensure efficient, quality resolutions to tickets. Plus, they can now guarantee consistency when it comes to issue resolution and improving SLAs, increasing the quality of their offer in the market and boosting their competitiveness.
TNBA works on all of MetTel’s tickets at all times, freeing up its operators and engineers to work on more complex tasks and resolving many of its tickets instantly.