Using AI/ML for Customer Service Chat Complexity to Improve Customer Experience

Image
Case Study: AI ML Solutions / Advanced Analytics

Using AI/ML for Customer Service Chat Complexity to Improve Customer Experience

SUMMARY

One of the largest telecommunications provider partnered with Xtrictech 3PL to incorporate artificial intelligence and machine learning to create a predictive model-driven flow to replace its rule-based online chat support system to improve the chat's success rate. Xtrictech 3PL 's solution achieved a success rate of 39% within the first six months of deployment.

Challenges

  • Client uses 3rd party system to manage the flow within the chat session on its website. The flow involves transfers among options, from automated billing info display to live agents manning the chat session. Rules drive the flow within each session.
  • The client wants predictive model-driven flow to replace its rule-based system to improve the chat’s success rate from 12% to 35%.
  • The success rate is defined as the average handling time of conversations at least 15% shorter than the AHT of corresponding calls.

Solutions Offered

  • The AI/ML solution estimates a chat session complexity composite index in real-time to predict the best flow pattern for the chat session. The combined complexity index consists of 10 components where the solution indicates the weight of each element to drive the following best flow action during the chat session.
  • The answer completely replaced the rule-based system, achieving a success rate of 39% within the first six months of deployment.
  • The deployment of the solution also improved customer experience and reduced the use of live agents overall, which resulted in significant cost savings.

RESULTS

0%
Success rate of chats within six months of deployment

Significant Reduction

In use of live agents overall — resulting in cost savings