Source
Which connected system supplied the information?
ChurnAI documentation
Connect a customer-data source, set up ChurnAI in your AI workspace, and investigate an account your team knows.
Start with one question. Check the evidence before expanding to your portfolio.
Choose an account with recent activity you can verify. This gives you a practical way to check the connection, review the available information, and assess the first answer.
The Pilot connects Chargebee or Amplitude. Choose billing or product usage for your first question. During onboarding, review the permissions and data included in the connection.
Confirm that ChurnAI associates the imported information with the correct customer. Check a familiar account before relying on broader findings.
Use the connection details provided during onboarding. Compatibility and authentication requirements vary by provider and plan.
Ask about the account, then compare the supporting evidence with the original source.
Replace the account name below with a customer you recognise.
Summarise what we know about [account name]. Include the sources and dates, and identify any missing information.
Follow up with:
Which changes deserve attention, and what evidence supports that assessment?
These questions make it easier to evaluate an answer before using it in a customer conversation.
Use these questions as starting points. Available answers depend on the sources and data connected to your workspace.
| Your task | Example question |
|---|---|
| Prepare for a renewal | What should I review before discussing renewal with [account]? |
| Investigate adoption | Where has usage changed for [account], and over what period? |
| Review an opportunity | What evidence suggests [account] could benefit from an expansion conversation? |
| Prepare a meeting | Create a meeting brief for [account], with recent changes and unresolved questions. |
| Review your portfolio | Which accounts need attention, and what evidence supports their priority? |
Which connected system supplied the information?
How recent is the evidence?
Does it refer to the correct customer?
Is relevant information missing?
If the evidence is incomplete, narrow the question or check the original system before using the answer.
MCP, the Model Context Protocol, lets a compatible AI workspace request customer context from ChurnAI.
Tell us whether you use Claude, ChatGPT, Codex, Cursor, or another workspace. We confirm compatibility during onboarding because MCP support, authentication, and data policies differ between providers.
Contact us for the endpoint and authentication instructions for your deployment. We’ll help you verify the connection using a known account before you begin broader portfolio questions.
Request workspace setup →During the Pilot, imported data is hosted in ChurnAI Cloud. From Pro onward, the current product offering uses customer-controlled storage.
Your AI workspace may process the context included in a request under its own data terms. Review the storage and access details for your plan before connecting customer information.
Confirm the permissions, imported fields, account identifiers, refresh schedule, and disconnection process for each source. The plan and integration determine which settings apply.
Compare plans and data options →Check that you’re using the connection details supplied for your deployment and that your workspace supports the required MCP connection. If authentication fails, send support the error text and workspace name.
Check whether the customer exists in the connected source and whether its identifier matches the expected profile. Ask support to review the mapping before relying on findings for that account.
Compare the evidence dates with the original system and the refresh schedule agreed during setup. Include the source name and expected update time when reporting a discrepancy.
Ask which sources and dates support the finding, and what information is missing. Verify that the relevant system is connected before using the answer in a customer conversation.
Support
Email us with your AI workspace, the integration you’re connecting, and the error you’re seeing. Omit passwords, API keys, and customer records.