Frequently Asked Questions
Will data remain private?
Yes. Informer's AI runs through Informer's cloud AI service using the providers' APIs, not a public consumer chatbot. Your data is not used to train the underlying models.
Why is the Assistant giving incorrect answers?
Sometimes, when the Assistant does not know an answer, it makes one up: this is called a hallucination. Always verify the Assistant's answers:
- UI verification: Compare what the Assistant says to what is on the screen. For example, a total it reports should match the aggregate in the grid.
- Explain: Each answer has a ? icon that asks the Assistant to explain how it reached its answer, giving transparency into its logic.
- Debug mode: Get a full technical view of the back-and-forth before the final answer.
Then adjust your prompt to get better responses.
Why is the Assistant not responding to questions?
The Assistant can get caught in a conversational loop. Enable Debug mode to check; if so, click the Stop icon to cancel the request, then rephrase the prompt or start over.
Why is the Assistant being slow or unresponsive?
The AI provider may be experiencing capacity issues or an outage. Check the provider's status page for current incidents.
Why is the Assistant responding with "message too long"?
Assistant conversations are limited to a maximum number of characters. Sometimes the Assistant requests too much data from a Dataset when it should use an aggregate query instead. Adjust the prompt to encourage an aggregation.
Why is the Assistant showing Python messages?
The Assistant occasionally tries to run a Python script even though Python is not supported. When this happens, click Start Over and rephrase the prompt.
Will Users have access to data they shouldn't?
No. All requests an Assistant makes are on behalf of the logged-in User, and that User's security and permissions apply to every interaction.