What challenges did DoorDash face before AI implementation?

Dashers and platform users need efficient self-service support to reduce wait times and live agent dependency. Phone support requires low response latency and accurate issue resolution to keep delivery workflows smooth and satisfactory for Dashers. DoorDash aimed to empower Dashers with reliable AI-driven phone support for common questions to improve user trust and productivity.

Source

ProductionEvidence: Medium55/100

DoorDash Builds Generative AI Contact Center Solution Using Amazon Bedrock and Amazon Connect Customer

Use case typeCustomer personalizationUpdated Jun 13, 2026

DoorDash, a local commerce platform, wanted to enhance self-service support for Dashers, Consumers, and Merchants by reducing live agent interactions and improving contact center user experience. Collaborating with AWS Generative AI Innovation Center, DoorDash built a voice-operated generative AI self-service contact center solution using Amazon Bedrock foundation models (Anthropic Claude) and Amazon Connect Customer, implementing retrieval-augmented generation from public help center data for accurate response. The AI-powered contact center handles hundreds of thousands of calls daily, reducing agent transfers by 49%, improving first contact resolution by 12%, and cutting development time by 50%. The solution operates with response latency of 2.5 seconds or less, is fully rolled out to all Dashers, and achieves $3M annual operational cost savings with plans for expansion.

Organization: DoorDash
Industry: Logistics
Location: United States
Published: May 2026

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: DoorDash
Provider: AWS
Maturity: Production
Linked source: AWS Customer Stories

5 seconds or less, is fully rolled out to all Dashers, and achieves $3M annual operational cost savings with plans for expansion

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 3 of 3

1Generative AI Contact Center
2Retrieval-Augmented Generation
3Self-Service Automation

DoorDash used Amazon Bedrock to access Anthropic Claude foundation models that mitigate hallucinations and abusive language with fast response times.
It deployed a retrieval-augmented generation approach leveraging Knowledge Bases for Amazon Bedrock to index public help center content for enriched AI responses.
Amazon Connect Customer provided the AI-native contact center platform for interactive voice response, integrated with generative AI for self-service.
Amazon SageMaker was used to build automated test and evaluation frameworks to reliably A/B test and scale the generative AI contact center solution.

Technologies

Amazon Bedrock Anthropic Claude Amazon Connect Customer Amazon SageMaker

Architecture

The solution architecture includes Amazon Bedrock foundation models (Anthropic Claude) for conversational AI, Knowledge Bases for Amazon Bedrock for retrieval-augmented generation, and Amazon Connect Customer as the AI-native voice contact center platform. It integrates automated testing via Amazon SageMaker for evaluation and scaling.

Implementation partners1

AWS Generative AI Innovation Center

Sources & evidence1

Evidence: Medium55/100Evidence strength