Phase 2: Tuning
Ensure your agents apply knowledge correctly
Once your AI has been taught, the next step is tuning. This phase is about refining and optimizing AI responses based on real-world scenarios.
Sandbox – A dedicated testing space where you can act like a guest, ask your AI questions, and adjust the underlying knowledge blocks to improve accuracy.
Inbox Proposed Replies – Reviewing AI-generated responses in live conversations, modifying sources when needed, and adding new knowledge directly from the inbox.
Advanced Rules – Defining custom rules that guide how the AI responds in more complex situations
Tuning helps you bridge the gap between knowledge and application, making sure your AI is not just informed, but responding the way you want it to in every interaction.
Related Docs
Last updated
Was this helpful?