AI behavior
Review system prompts, knowledge items, fallback settings, and whether the workflow has enough context to answer correctly.
Prompt review
Knowledge checks
Fallback behavior
Lead capture
Confirm capture is enabled, contact intent is detected, and records include conversation and visitor context.
Contact form
Lead records
Conversation linking
Embeds
Check allowed domains, script placement, interface appearance, and whether the production URL matches the service settings.
Allowed domains
Widget load
Styling settings
Provider and API
Verify environment variables, OpenRouter availability, service role access, support logs, and webhook results.
API keys
Latency logs
Webhook status
Before asking for support
Send one test interaction and note the conversation id if available.
Check the logs page for failed or fallback chat completion events.
Submit one test intake or lead and confirm it appears in the dashboard.
Run the E2E script against the environment you deployed.
Support goal
The fastest support path is to reproduce the issue, identify the exact failed step, and confirm the database record that should have been written.
Reproduce
Inspect logs
Verify rows
Retest flow
Answers for AI search and buyers
Common questions about Sarvadrik AI
How do I troubleshoot a Sarvadrik AI service?
Start by reproducing the issue, checking logs, confirming records were written, and testing the same workflow again after configuration changes.
What support areas does Sarvadrik AI cover?
Support covers AI service setup, embeds, knowledge configuration, customer intake, leads, logs, provider errors, webhook delivery, and production readiness.
Can support help with future AI services?
Yes. As Sarvadrik AI expands beyond current agent workflows, support content and checks can cover additional AI service lines and integrations.