Customer support · Production

Enhancing LLM accuracy with Coveo Passage Retrieval on Amazon Bedrock

The problem

Enterprises adopting LLMs face the challenge that without reliable data foundations these models can generate misleading or inaccurate responses, reducing user trust and organizational credibility. In RAG systems, extracting the most relevant, precise information from enterprise data sources is the most complex component.

Workflow diagram · grounded in source

User submits natural language query

trigger

“Amazon Bedrock Agents acts as the intelligent backbone of this solution, interpreting natural language queries and orchestrating the retrieval process to deliver grounded, contextually relevant insights”

Agent invokes Passage Retrieval API

integration

“it invokes the CoveoPRAP action group, which is specifically designed to retrieve relevant passages, through a Lambda function to make an API call to /rest/search/v3/passages/retrieve”

Two-stage passage extraction

ai_action

“In the first retrieval stage, Coveo's hybrid search system is used to identify the most relevant documents. Then, it extracts the most relevant text passages from these documents, along with ranking scores, citation links, and other key …”

ML-driven relevancy optimization

ai_action

“AI continuously learns from user interactions, tailoring retrieval to each user's journey, behavior, and profile for context-aware responses”

LLM generates grounded response

output

“generated answers are grounded in an organization's proprietary knowledge”

Analytics track answer performance

feedback_loop

“With events tracking through the Data Platform and Knowledge Hub, you can see exactly how your generated answers perform, where information is missing or underused, and which content needs tuning”

Reported outcome

By integrating Coveo's Passage Retrieval API with Amazon Bedrock Agents, organizations can develop AI applications that provide validated responses based on enterprise content, helping reduce inaccuracies while delivering real-time, secure responses.

Reported metrics

LLM response inaccuraciesreduce inaccuracies

Reported stack

Coveo AI-Relevance PlatformAmazon Bedrock AgentsAWS LambdaAmazon CloudWatchAWS CloudFormationClaude 3 Haiku v1SalesforceSharePointGoogle Drive

Source

https://aws.amazon.com/blogs/machine-learning/enhancing-llm-accuracy-with-coveo-passage-retrieval-on-amazon-bedrock?tag=soumet-20

Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

By integrating Coveo's Passage Retrieval API with Amazon Bedrock Agents, organizations can develop AI applications that provide validated responses based on enterprise content, helping reduce inaccuracies while delive…

What tools did this team use?

Coveo AI-Relevance Platform, Amazon Bedrock Agents, AWS Lambda, Amazon CloudWatch, AWS CloudFormation, Claude 3 Haiku v1, Salesforce, SharePoint, Google Drive.

What results were reported?

LLM response inaccuracies: reduce inaccuracies (source-reported, not independently verified).

How is this customer support AI workflow structured?

User submits natural language query → Agent invokes Passage Retrieval API → Two-stage passage extraction → ML-driven relevancy optimization → LLM generates grounded response → Analytics track answer performance.