Context filtering in Genesys Agent Copilot

AI-generated summary

Genesys Cloud's Agent Copilot uses context filtering to dynamically surface relevant knowledge articles during agent interactions by leveraging active intents and conversation custom attributes. Administrators can create custom context attributes that automatically filter knowledge article presentation to agents, with Agent Copilot extracting context values from conversation attributes and active intents during interactions. Implementation requires several configuration steps: defining conversation schemas to establish data structure and type (supporting string attributes only), updating conversation attributes via PATCH or PUT API endpoints, configuring data actions through the Genesys Cloud Platform API with PUT method and a 60-second execution timeout, and setting conversation custom attributes using Architect flows or Scripter. The data action request body must include id (arbitrary UUID), organizationId, schemaId, product, and brand variables. Additional setup involves querying the conversation attribute API (GET endpoint) in Scripter to retrieve initial values and wire them to variables or input fields for dynamic updates. Finally, administrators must configure knowledge articles with appropriate labels in the Knowledge Workbench and establish context filters within Agent Copilot settings to complete the implementation.

Genesys Agent Copilot can use the context to filter the knowledge it surfaces during conversations. It processes contextual data from active intents and conversation custom attributes.

Administrators can create custom context attributes that automatically filter the knowledge articles Agent Copilot presents to agents. Agent Copilot extracts context source values during conversation from conversation attributes and active intent. Active intent is always part of the context after you define the intents.

To use context filtering, you must perform the following setup steps.

Define conversation schemas. These schemas determine the data structure and type. For more information, see Conversation schemas overview.
Note: Agent Copilot only supports string attributes.
Update conversation attributes through the API. During the conversation, you can update the conversation attributes through the following APIs.
- PATCH /api/v2/conversations/{conversationId}/customattributes
- PUT /api/v2/conversations/{conversationId}/customattributes
Configure data actions. Data actions provide a way to interact with runtime data through the Genesys Cloud Platform API. For more information, see About the Genesys Cloud data actions integration.
1. In the HTTP Method list, select PUT.
2. In the Request URL Template field, enter /api/v2/conversations/{conversationId}/customattributes.
3. In the Execution Timeout field, enter the value 60.
4. In the Request Body Template field, enter the following variables:
  - id: an arbitrary UUID
  - organizationId
  - schemaId
  - product (from Architect or Scripter)
  - brand (from Architect or Scripter)
Set conversation custom attributes using an Architect flow, using the data action created in the previous step. For more information, see Use data actions in Architect for integrations.
Set up data actions in Scripter. For more information, see Available script actions.
When the script page loads, you need a data action to query the conversation attribute API. The API is GET /api/v2/conversations/{conversationId}/customattributes. When the initial state query is complete, then you must wire the value to a variable or input field.
You can change the initial value through an input field or a drop-down, and connect the data that updates the conversation attribute when a change occurs.
Configure knowledge articles and labels. For more information, see Knowledge workbench overview V2 and Add labels to a knowledge base V2.
Set up context and add context filters to the knowledge base for Agent Copilot. For more information, see Configure Agent Copilot settings.

[NEXT] Was this article helpful?

Get user feedback about articles.