What was the impact of the AI solution?

The blog says the assistant can reduce mean time to resolution. It also says the solution can improve first-time fix rates by helping technicians arrive with the right parts and procedures. The article emphasizes grounded diagnostic recommendations from manufacturer documentation and a single AgentCore Runtime endpoint that simplifies operations.

Source

AI-Powered Equipment Repair Assistant Using Amazon Bedrock AgentCore (John Deere documentation-based demo)

Use case typeWorkflow automationUpdated Jun 10, 2026

This AWS blog shows how to build an AI-powered equipment repair assistant for heavy farm machinery using Amazon Bedrock AgentCore. The assistant helps technicians diagnose problems, identify parts, and access manufacturer-approved repair procedures through natural language. The solution uses Amazon Bedrock AgentCore Runtime with the Strands Agents SDK, Amazon Bedrock Knowledge Bases for retrieval-augmented generation, Amazon Nova 2 Lite, and AgentCore Memory for conversation persistence.

Organization: John Deere
Industry: Agriculture
Location: United States
Published: June 2026

Reported outcomes

Strategic outcomes

Cost efficiencyReduced mean time to resolutionOther strategic outcomeImproved first-time fix rates

Primary read

Use case focus

Showing 1 of 1

1Workflow automation

Technicians must diagnose heavy farm machinery issues without the right parts or procedures, which can cause multiple site visits, extended downtime, and financial losses during harvest season.

A web app with Amazon Cognito authentication calls an AgentCore Runtime endpoint.
The agent uses a custom search tool that queries an Amazon Bedrock Knowledge Base built from equipment manuals, parts catalogs, and repair documentation stored in Amazon S3 with vector search in Amazon OpenSearch Serverless and Amazon Titan embeddings.
Amazon DynamoDB stores service tickets and AgentCore Memory preserves short-term and long-term conversation context.
Amazon CloudWatch and AWS X-Ray provide observability for the agent runtime.

Technologies

Amazon Bedrock AgentCore Amazon Bedrock Knowledge Bases Amazon Nova Lite Amazon Cognito AWS Amplify Amazon S3 Amazon OpenSearch Serverless Amazon Titan embeddings Amazon DynamoDB Amazon CloudWatch AWS X-Ray

The blog says the assistant can reduce mean time to resolution.
It also says the solution can improve first-time fix rates by helping technicians arrive with the right parts and procedures.
The article emphasizes grounded diagnostic recommendations from manufacturer documentation and a single AgentCore Runtime endpoint that simplifies operations.

Architecture

A CloudFormation-deployed AWS architecture combines Amazon Cognito and AWS Amplify for a React frontend, an Amazon Bedrock AgentCore Runtime endpoint hosted with the Strands Agents SDK, a custom tool that invokes Amazon Bedrock Knowledge Bases retrieval-and-generation against equipment documentation in Amazon S3 and Amazon OpenSearch Serverless with Amazon Titan embeddings, Amazon Nova 2 Lite for inference, AgentCore Memory for state persistence, Amazon DynamoDB for issue CRUD, and Amazon CloudWatch plus AWS X-Ray for observability.

Sources & evidence1

Groundedness: 4/5Type: Blog PostPublished: Jun 10, 2026Publisher: AWSEvidence: VendorConfidence: Medium

Primary source

AI-generated summary. Verify important details with the linked sources before relying on this case.

Explore related AI use cases

Customer evidence