Implement artificial intelligence agents in your customer service center for voice and text with Amazon Connect, Amazon Lex, and Amazon Bedrock Knowledge Bases | AWS Machine Learning Blog

The Challenge Faced by DoorDash

DoorDash, a company connecting consumers with local businesses globally, struggled to manage the high volume of calls from its contractor delivery workers known as Dashers. With millions of active consumers and Dashers, the need for a more efficient self-service experience became crucial.

Solution Development with Generative AI

DoorDash collaborated with the AWS Generative AI Innovation Center to implement a voice-enabled self-service solution for Dashers. This solution utilized services like Amazon Lex, Amazon Bedrock, and Amazon Bedrock Knowledge Bases to provide quick and reliable assistance to Dashers.

Benefits and Operational Capabilities

The implemented solution successfully reduced the reliance on live agents, with the ability to handle hundreds of thousands of calls daily and respond to Dashers in 2.5 seconds or less. It also offered operational features such as automated testing, conversation analytics, and hallucination prevention and detection.

Deployment and Testing of the Solution

Steps were outlined for deploying the generative AI agents in a contact center using AWS services, with a sample interaction showcased between a customer and the LLM-powered automated agent. The process involved the creation of a CloudFormation stack and setting up the necessary resources for the application.

Utilizing QuickSight for Analytics

The solution leveraged QuickSight for analytics, with a detailed architecture diagram illustrating how conversation log entries were processed and visualized in near real-time dashboards. The steps for setting up dashboards and visualizations in QuickSight were also provided.

Automated Testing and Solution Adaptation

Processes for trying the automated testing capability and adapting the solution to specific use cases were explained, emphasizing the scalable and cost-effective nature of the contact center generative AI agent solution using AWS services.

Conclusion

The article highlighted the successful collaboration between DoorDash and AWS in implementing a voice-enabled self-service solution using generative AI. It encouraged readers to explore the open-source solution code and engage with AWS services for expert assistance and further development.