Skip to main content
Cymph AI is a conversational assistant built into the platform. It is grounded in the data you have access to: your playbooks and their governance, your assets, your framework mappings and your execution history. That context is what lets it go beyond generic advice and work on your actual procedures. The assistant can:
AI features require the AI settings of your instance to be configured. Nothing is sent to a model unless a user explicitly asks the assistant for something. See AI & data usage.

Opening Cymph AI

There are two ways to open the assistant:
  • Click Cymph AI from the left navigation menu.
Ai Assistant Nav
  • Click the Ask AI button available on most screens.
Ai Sidepanel

Drafting a new playbook

Describe the playbook you need, for example “I want a playbook for Windows”. Rather than guessing, the assistant asks the questions it needs to shape the draft, such as whether you want an incident response runbook for a specific alert, a high-level response plan or a generic technical playbook, and lets you describe exactly what it should cover. Answer by picking an option or typing.
Ai New Playbook
When the draft is ready, it appears as a card in the conversation. Click Review to open it in the panel on the right, where the full documentation is editable in place. Add labels, adjust the content, then click Save to Workspace. Before the draft is saved, Cymph asks for the Responsible and Accountable persons and a review frequency, so the new playbook starts with its governance context in place.
Ai Assisted Review
You can also start this flow from Playbooks → Library → New Playbook → Generate with AI. See Creating your first AI-assisted playbook.

Improving existing playbooks

Ask the assistant to improve a playbook, for example “Can you suggest improvements for the playbook RANSOM-RECOVERY?”, or describe what changed in your environment or what an exercise showed. The assistant uses the playbook itself and the context around it: its assets and governance, its framework mappings, and any executions with their improvements, skip and fail reasons and timings. The response is a set of proposed changes, not a rewritten playbook. The assistant summarises what it proposes, lists each affected playbook with the number of additions and removals, and marks it For Review. Nothing is changed until you decide. For each playbook you can:
  • Click Review change to inspect the proposal before applying it.
  • Click Apply update to apply it directly, when the summary is enough.

Reviewing a proposed change

Ai Updates Real
The review panel shows the proposal against the current content of the playbook. Removed text is struck through in red and new text is highlighted in green, with a running count of additions and removals at the top. Switch between the Documentation, Properties and Workflow tabs to see every part of the playbook the proposal touches.
  • Click Update & Next to apply the change and move to the next proposal.
  • Click Skip to leave the playbook as it is and move on.
  • Click Open in editor if you want to adjust the proposal by hand rather than accept or skip it as a whole.
When several playbooks are affected, the pager at the top of the panel shows where you are, and you can work through them one at a time. Applied changes are saved to the playbook like any other edit, so they appear in its history and are covered by the review settings you have in place.
Be specific in the request. “Update the isolation step so it meets the 10 minute target from the last exercise” produces a more targeted proposal than “improve this playbook”. See Turn operational context into better response for how this fits into a continuous improvement loop.

Asking questions

Beyond generating and improving playbooks, the assistant answers questions about your workspace, summarises playbooks and documents, and searches the sources connected to it. Answers are grounded in the content you have access to, so the assistant never reveals playbooks or assets that are not shared with you.

Chats

Conversations are saved as chats so you can return to them later.
  • Click New chat to start a fresh conversation.
  • Use the Search chats box at the top of the chat list to find a previous conversation.
  • Chats can be renamed via their inline menu action.
  • Chats can be deleted.

Attachments and voice

  • Click the + button to attach a file for additional context, for example a debrief document or an existing procedure.
  • Click the microphone icon to dictate your message instead of typing.