Goldilocks Docs
Getting Started

Launch Checklist

Pre-launch, testing, and go-live steps for your AI agent

Use this checklist before and after going live with your AI support agent.

Pre-Launch

Content Quality

  • Documents are up to date and accurate
  • FAQs cover common questions
  • No conflicting or contradictory information (check Quality Center)
  • Content is published (not draft)

Persona Configuration

  • Support persona has appropriate tone and instructions
  • Escalation settings are configured
  • Custom instructions reflect your policies

Escalation

  • Escalation email is correct
  • "Collect before escalating" fields are set if needed
  • Escalation workflow is tested
  • Team knows how to receive escalations

Chat Branding

  • Logo and colors match your brand
  • Greeting message is set
  • Position (e.g., bottom-right) is configured

Testing

Search Testing

  • Use Training > Search to test key questions
  • Verify responses are accurate and grounded in content
  • Check that irrelevant questions are handled gracefully

Chat Preview

  • Test in Installation > Preview
  • Try suggested questions
  • Test escalation flow
  • Verify Context configuration if used

Staging

  • Deploy to staging site if available
  • Test on mobile
  • Test across browsers
  • Verify domain allowlist includes staging URL

Go-Live

Embed Code

  • Copy embed code from Installation (click Get Embed Code in the Essentials section or header)
  • Add to your site (before </body>)
  • Use data-key attribute with your API key
  • Add persona slug to embed code if using a specific persona

Domain Allowlist

  • Add production domain(s) in Installation or Settings
  • Remove test domains if needed

Final Check

  • Load production site and open widget
  • Send a test message
  • Confirm AI responds correctly
  • Test escalation if applicable

Post-Launch

Monitor

  • Check Rundowns for conversation summaries
  • Review Contact Signals for topic trends
  • Monitor first escalations and respond promptly

Iterate

  • Add content based on unanswered questions
  • Refine persona instructions based on feedback
  • Update FAQs when policies change