Every prompt below runs against your live partner data through one secure toolset. Run it in Introw AI, the assistant built into Introw, or from your own assistant over Introw’s MCP server - Claude, ChatGPT, Gemini, Cursor. The surface changes, the prompt does not: the same tools, your own permissions, and every write landing in your CRM.
Where to run them
Introw AI runs the same toolset the MCP server exposes, so a prompt written for Claude works unchanged inside Introw, and the other way round. Partners get the same choice: their own assistant through Partner AI, or the support agent that answers them in the portal, Slack, Teams, email and WhatsApp.How to use this page
1
Pick your surface
Introw AI is already there: open it from the top of Introw and ask. To use your own assistant, connect it once over MCP - see Connect Claude, Connect ChatGPT, or Connect an MCP client for Gemini, Cursor and the rest.
2
Copy a prompt and edit the specifics
Hover a prompt and use the copy button. Swap the partner names, tiers, and time ranges for your own; the phrasing does not have to be exact, the assistant matches on intent.
3
Approve anything that writes
Reads run straight away. Actions that change data follow your capability matrix, so they either execute, wait for approval, or stay blocked. See Governance and trust.
Prompts are grouped by the product area they act on, and each block links to the page that documents the capability behind it. If a prompt returns nothing, the usual cause is that the feature is not configured yet or your role cannot see the data, not that the assistant misunderstood.