AWS connects Bedrock Data Automation and Model Context Protocol with Salesforce Agentforce for the public sector
AWS described an architecture connecting Amazon Bedrock Data Automation and Model Context Protocol with Salesforce Agentforce – it automatically extracts insights from video, photographs and documents for the public sector and makes them accessible through natural language queries in Salesforce.
AWS described an architecture in a technical blog post that combines Amazon Bedrock Data Automation with Model Context Protocol (MCP) and enables Salesforce Agentforce to automatically process unstructured data – body camera video, surveillance footage, photographs and scanned documents – in the public sector. The solution builds on an earlier integration for storing files in Amazon S3 within Salesforce Public Sector Solutions and addresses the second part of the problem: automatically extracting insights from stored data.
The architecture uses event-driven processing: uploading a file to Amazon S3 triggers an AWS Lambda function, which creates a document identifier, stores metadata in Amazon DynamoDB and starts processing in Amazon Bedrock Data Automation. Depending on the media type, the service extracts text and key fields from documents, descriptions and recognized objects from images, or transcripts and scene summaries from video and audio. The results are stored in an S3 output bucket and made available for further querying through Amazon EventBridge.
On the Salesforce side, a user query in Agentforce triggers a call via MCP that passes through Amazon Bedrock AgentCore Gateway – which authenticates the request and forwards it to an MCP server running on AWS Lambda. The server retrieves matching records from DynamoDB and S3 and returns them to the agent context, so the user receives a natural language response without leaving the Salesforce console. AWS states that the architecture is modular and can be extended beyond evidence management, and provides sample deployable code (an AWS CDK stack) on GitHub, which, according to its own statement, is not intended for direct production use.
Registration of external MCP servers is available in Salesforce Developer, Enterprise, Performance and Unlimited Edition. The source article also describes specific deployment steps, but their full text was not available. You can find details in the source article.
Why it matters
For public sector agencies and integrators working with Salesforce and AWS, this provides a concrete architectural pattern for reducing manual review and sorting of evidence materials – analysts can search for insights in natural language directly in Salesforce instead of switching between systems. However, deployment requires an existing integration between Salesforce Public Sector Solutions and Amazon S3 and an appropriate Salesforce edition that supports registration of external MCP servers; according to AWS, the provided code is only a sample and requires users to implement their own security measures before production use.
Relevant practical impact
What this means
For a business
Public sector organizations using Salesforce Agentforce Public Sector and AWS can automate the extraction of insights from unstructured evidence materials (body camera video, photographs, documents) and search for them using natural language queries directly in Salesforce, eliminating manual data sorting; according to AWS, however, this is sample code that is not intended for direct…
ProcessesCheck the original
Event sources
only one source so far · 1 publisher, 0 independent. We count feeds from the same owner only once.