Search Agent implementation overview

This is for:

Developer System Administrator
Important
Beta feature

Coveo Search Agent is currently available as a beta offering. Contact your Customer Success Manager for access to this feature. Your use of this feature is subject to the beta and pre-release terms of your agreement with Coveo, including any applicable beta or pre-release provisions therein. To the extent your agreement does not contain specific beta or pre-release terms, Section 8 (Beta Features) of the Coveo Customer Agreement shall apply. This feature is provided "as-is," without warranty or SLA coverage, and may be modified, suspended, or discontinued at any time. You should not use this feature to process sensitive or regulated data.

This article outlines the steps required to implement a Coveo Search Agent in a Coveo-powered search interface.

Step 1: Choose the content to use

Step 2: Create a CPR model

Step 3: Create a Semantic Encoder (SE) model

Step 4: Associate the SE model with your query pipeline

Step 5: Create the Search Agent

Step 6: Configure a search interface for the Search Agent

Note

The answer generation model that’s used by the Search Agent is created and managed by Coveo personnel.

Step 1: Choose the content to use

When using the Search Agent to generate answers in a Coveo-powered search interface, the Passage Retrieval (CPR) model is responsible for retrieving the segments of text (passages) from which the answers are generated. The content that you choose to use for the CPR model determines the raw data from which passages are retrieved and answers are generated.

See Passage Retrieval (CPR) content requirements and best practices.

Note

A Passage Retrieval (CPR) implementation must include both a CPR model and a Semantic Encoder (SE) model. Both the CPR and SE models must be configured to use the same content.

Step 2: Create a CPR model

Create the CPR model that the Search Agent will use to retrieve passages from which to generate answers.

See Create and manage Passage Retrieval (CPR) models.

Note

You don’t need to associate the CPR model with your search interface’s query pipeline as the Search Agent calls the CPR model directly and not through the query pipeline.

Step 3: Create a Semantic Encoder (SE) model

The SE model adds vector-based search capabilities to your Coveo-powered search interface, enabling it to find items based on semantic similarity with the query. To ensure answers are generated from the most relevant content, the SE model, as part of the query pipeline, helps to identify the most relevant items from the index.

See Create and manage Semantic Encoder (SE) models.

Note

A Passage Retrieval (CPR) implementation must include both a CPR model and a Semantic Encoder (SE) model. Both the CPR and SE models must be configured to use the same content.

Step 4: Associate the SE model with your query pipeline

Associate the SE model with the query pipeline that’s used by the Coveo-powered search interface that will use the Search Agent.

See Associate a Semantic Encoder (SE) model with a query pipeline.

Step 5: Create the Search Agent

Create the Search Agent configuration for your search interface.

See Manage Coveo Search Agents.

Step 6: Configure a search interface for the Search Agent

Add the Search Agent to your Coveo-powered search interface to enable agentic conversational search and answer generation.

See Configure a search interface for the Search Agent.