Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
Knowledge sources ground your agent in real data, enabling it to answer questions using your organization's content rather than relying solely on its general training. In agents powered by the GitHub Copilot harness, you configure knowledge from the Build tab.
Note
Features in this article are powered by the GitHub Copilot harness, which uses Copilot Credits for usage-based billing. Learn more about Copilot Credits in Overview of billing for agents powered by the GitHub Copilot harness.
Usage-based billing applies to using, building, testing, and evaluating agents. These actions might consume Copilot Credits.
What is knowledge?
Knowledge sources are the data your agent can access when answering user questions. Without knowledge, the agent relies only on its general training and instructions. With knowledge, the agent can reference specific documents, websites, organizational data, and structured databases.
The enhanced orchestration runtime automatically determines which knowledge sources to search based on the user's question.
Why add knowledge to your agent?
- Ground responses in specific, accurate information
- Keep answers current with your organization's latest content
- Scope the agent to answer only from approved sources
- Allow the agent to reference internal documents, policies, and procedures
How the agent uses knowledge
- When a user asks a question, the orchestration runtime evaluates whether knowledge sources are needed
- The agent searches relevant sources, retrieves content, and uses it to formulate a response
- Citations might be included to show where information came from
- The agent's instructions influence how it interprets and presents knowledge