How a US logistics enterprise reduced its customer query resolution time by 95%

Our client uses Freshdesk to manage the customer support tickets. With a growing customer base, they began receiving a large volume of tickets. 

Most of these tickets were repetitive, with similar issues being raised by different clients. Resolving these took most of the bandwidth of the support team. Hence, it was taking on average ~2-3 hours (AHT) to resolve a ticket. Additionally, the priority tickets were also getting queued for a long time.

We deployed our AI powered agents and customized it based on the client requirements. The AI agents were integrated with FreshDesk, client’s customer ticketing system.

Our AI agent automatically resolved the repetitive tickets without any manual intervention. About 15% of the resolutions were sent for human verification before responding to the customer.

The process included following 3 steps: 

‍

Step 1: Intent Identification from customer query

An intent identification AI agent was created. It was trained on 30k+ historical tickets of the client. Each new incoming ticket was passed to the agent to figure out the intent of the Customer query. (eg. “AWB not generated, User creation failed”)

‍

Step 2: Automated query resolution

Each intent has an AI agent to handle the query, debug it and generate the appropriate response. Within each AI agent, the relevant information was retrieved and actions performed across a variety of platforms used in the Logistics industry.

At the end of the workflow, an appropriate response is generated and sent back to the customer.

‍

Step 3: Query escalation

~15% tickets that scored low on AI confidence level were escalated for human verification before responding to the customer. 

‍