Run with AI turns current In Parallel context into a focused handoff for an AI tool. When you have selected a preferred client, the button can name it directly, such as Run with Claude Code.
Where Run with AI appears
The current product shows a Run handoff in these contexts:
Workspace pages — continue from the content of a Charter, Runbook, Retro Log, or another maintained page.
Meeting Actions — continue from an AI-drafted follow-up, such as an email recap, tracker update, or Slack post.
Some settings and accessible copy may still use the earlier Send to AI label. Follow the visible button label on the page or meeting Action you are working from.
Run a Workspace page with AI
Open the maintained page you want to work from.
Select Run with [preferred client].
Review the handoff in the receiving tool before you continue.
Open the adjacent menu and choose Copy as prompt when you prefer to paste the prepared handoff yourself, or Choose every time… to select a client for each handoff.
A page handoff is useful for work such as:
Turning a Runbook into a checklist for a specific incident
Preparing a handover from the current Charter and operating context
Reviewing objectives for gaps or contradictions
Drafting a decision brief from a maintained page
Run a meeting follow-up with AI
Open the meeting and select Actions. When In Parallel has prepared follow-ups, review the proposed task before running it.
The follow-up is a starting point, not a completed external action. Check the audience, destination, and wording in the receiving workflow before anything is sent or saved.
Run with AI or Chat?
Use | When |
Chat | You want to ask a question or explore context inside In Parallel |
Run with AI/client | You want to continue a specific Workspace page or meeting Action in another AI workflow |
Connected assistant over MCP | You want an external assistant to retrieve authorized context across multiple In Parallel records |
Privacy and review
The handoff goes to the AI tool you choose, which has its own privacy and retention policy. Review the prepared context before continuing, especially when the page or meeting contains sensitive information.
The receiving AI can still make mistakes. Check important claims against the source record and keep a person responsible for consequential decisions, communications, and writes.