2. Click on the chat icon next to your assistant.#
This opens the detailed configuration view for that specific assistant.
3. Here you can see whether your assistant is currently in the Production environment or the Staging environment.#
4. If your assistant is in the Production environment, all these options will be disabled.#
These options are read-only in the Production environment.
5. If you want to change the assistant’s configuration, switch to the Staging environment.#
All edits to the assistant's configuration should be done in Staging before publishing to Production.
6. Certain user-facing actions like Share (generate links), Like/Dislike (vote), Add to Favorites, and Submit Feedback are disabled in staging because they only apply to the live production agent#
7. All these options are now enabled, and you can update your assistant.#
Review and configure which tools are enabled — these define what the assistant can access or interact with during a conversation.
10. Click on Corpus#
11. Click the Add New (➕) icon at the top-right of the Corpus panel.#
This initiates the process of attaching a new knowledge corpus to the assistant.
12. A modal/dialog appears showing available corpora. Select the desired corpus from the list to attach it to the assistant.#
13. Click on Guardrails#
Guardrails control what topics the assistant is allowed or restricted from discussing — this is critical for ensuring the assistant stays on-topic and within company policy.
14. Click on Customize to access role configuration for the assistant.#
15. Here you can change your assistant’s role.#
16. Here you can set custom instructions and conversation starters for your assistant.#
17. Click on Advanced Section#
18. In the Advanced panel, a model selector is visible. click it to choose which large language model the assistant should use.#
19. Here you can select the reasoning pattern for your assistant.#
20. Once you’ve made all your changes in the Staging environment, Click the Deploy button at the top of the page.#
21. A confirmation dialog appears with a summary of the changes being published.#
Review the details shown such as model, guardrails, or corpus updates — to verify everything is correct before proceeding.
22. Write a short description of the changes you made.#
23. Click Deploy Changes to confirm and push the assistant configuration live to Production.#
24. The view switches back to the Production tab automatically after a successful publish.#
You can now see the live state of your assistant as it is deployed to end users.
25. Click the Staging tab to navigate back to the Staging environment for any future iterations or updates.#
26. Click on Deployment History in the right panel to view a log of all past publications for this assistant.#
This gives you a versioned record of every time changes were pushed to Production, allowing for auditing and rollback if needed.
27. Click on a past deployment entry (e.g., Deployment #2) to expand and view its specific details including the timestamp, configurations, and what changed.#
Use this when a recent deployment introduces unexpected issues — restoring instantly restores the assistant to its prior working state.
30. Click Deploy Changes to confirm and push the assistant configuration live to Production.#
