Tips for a successful Go Live

​

Before go-live (prep that pays off)

• Ensure your team has completed the testing phase of onboarding
  • This includes filling out the Testing Feedback document that your CSM has shared with you during your testing call.
  • Make sure all of the updates to your bot have been completed by the Avoca team. Your CSM should inform you of this but feel free to follow up prior to your Go Live call
• Make sure your team is ready for the new workflow. Share a quick internal heads-up:
  • What changes on go-live day
  • Who to contact if something looks wrong
    • Email your CSM or support@avoca.ai
  • How to report calls (LINK to other article)
• Make sure everyone on your team who wants to receive notifications has enabled that
  1. (LINK)
• Make sure you have configured call reasons
  • (LINK)

​

Day-of go-live (what to do during the first calls)

• Expect adjustments given real customer data: flag to your CSM any calls (by sending the call link as a reference) that you have questions about
• Listen to the first real calls. Spot-check a handful of early interactions to ensure:
  • The assistant is understanding callers
  • The right information is being collected
  • Transfers (if any) are landing correctly
  • Bookings or follow-ups are happening as expected

​

First week (how to improve quickly without chaos)

• Review patterns daily. A short daily check helps you catch issues early:
  • Calls that didn’t complete as expected
  • Common caller questions the assistant should address
  • Any categories that should be re-labeled or routed differently
• Tune the “top 3” issues first. Small changes to high-volume scenarios usually create the biggest impact.

​

Common go-live pitfalls (and how to avoid them)

• Pitfall: Misaligned hours or after-hours behavior.
  • Fix: Confirm hours, holidays, and emergency messaging before go-live.
• Pitfall: Inconsistent service naming (what customers say vs. what your system expects).
  • Fix: Add common synonyms and examples to reduce confusion.
• Pitfall: Too many new behaviors at once.
  • Fix: Launch the core call flow first, then iterate. Keep it simple.