Common issues
Most issues fall into a handful of patterns. Before you open a support ticket, run through the checks below.
Faster route for many of these: open the failing conversation, then ask Studio Assistant to fix it — “the agent told the caller we offer same-day delivery; we don’t, update the shipping topic”. See Iterate after a call for the full loop.
Quick reference
Knowledge base issues
Agent gives a wrong or unexpected answer
Diagnosis steps:
- Open the conversation in Conversations and enable Topic citations in the Diagnosis toggle group on the Transcription tab.
- Check which topic was retrieved – is it the right one?
- If the wrong topic was retrieved, the sample questions or topic names may be too similar between topics. Make them more distinct.
- If the right topic was retrieved but the answer is wrong, update the Content field.
Common causes:
- Two topics with overlapping sample questions – the retriever picks the wrong one
- Topic name is too vague (e.g.,
infoorgeneral) and matches too broadly - Content is outdated or incomplete
Agent says 'I don't know' for a topic that exists
Diagnosis steps:
- Check that the topic has been published and promoted to the environment you are testing in.
- Verify the topic has sample questions – without them, retrieval accuracy drops significantly.
- Test different phrasings. If only very specific wording triggers the topic, add more varied sample questions.
Common causes:
- Topic exists in Sandbox but hasn’t been promoted to Live
- Sample questions are too few or too narrow
- Topic name is not semantically related to the user’s query
Agent ignores the Actions section
When a topic instructs the agent to both say something and call a function in the same turn, the agent may do one but not the other. This is a common anti-pattern.
Fix: Structure actions so the agent only has one task per turn. If you need to say something and then call a function, split the instruction across turns – let the agent deliver the answer first, then handle the action in a follow-up.
Voice and audio issues
Voice speed change isn't reflected in calls
The change may be pulling from cached audio. Go to Voice > Audio library, find the specific audio clip, update the speed setting, and regenerate the audio.
Agent mispronounces a word or brand name
Use the Pronunciation section in Voice > Advanced settings > Speech. You can provide IPA notation, SSML overrides, or regex-based corrections. See Audio Management for details.
Disclaimer voice sounds different from the main agent
The disclaimer uses a separate voice configuration. Check Voice > Settings – the Disclaimer section has its own voice selection and tuning, independent of the main agent voice.
Deployment and version issues
Changes work in Sandbox but not in Live
You need to publish your changes and then promote them through your environments (Sandbox → Pre-release → Live). Changes in Draft or Sandbox are not automatically applied to Live.
See Environments for the full promotion flow.
I promoted but the old behavior persists
- Wait a few minutes – promotion can take a short time to propagate.
- Check that you promoted the correct version by comparing version numbers in the Environments page.
- If using a phone number to test, confirm the number is assigned to the correct environment.
Call drops after the disclaimer
If there are no knowledge topics published to the Live environment, the agent has nothing to say after the disclaimer and will end the call. Ensure at least one topic is published and promoted before going live.
Function and flow issues
Function isn't being called by the agent
- Check that the function is referenced in a topic’s Actions field or a step’s function list using
@function_name. - Verify the function description is clear – the LLM uses it to decide when to call the function.
- Check for typos in the function name reference.
- Use
conv.logstatements to confirm whether the function is being reached at all.
Flow skips a step or transitions incorrectly
Check the goto_step calls in your flow functions. A common bug is using multiple if statements instead of if/elif/else, which causes later transitions to overwrite earlier ones. See Flow fundamentals for the correct pattern.
Function returns an error
- Add
conv.logstatements to trace execution. - Check the function’s parameters match what the LLM is passing.
- If calling an external API, verify secrets are configured for the correct environment.
- Review the conversation in Conversations with the Tool calls diagnosis toggle enabled.
See Function maintenance for detailed debugging steps.
Telephony and call routing issues
Call isn't being forwarded to the agent
- Verify the forwarding number is correct and includes the country code.
- Check that call forwarding is enabled with your telephone provider.
- Ensure the transfer number is not part of the same ring group as the mainline number, which can create a loop.
Call transfers fail or drop
- Check that the transfer number includes the country code (starts with
+). - For SIP transfers, verify the URI and any custom headers.
- Test in Sandbox first before promoting to Live.
- If using Twilio, confirm the fallback number is configured correctly.
See Routing and handoffs for step-by-step guidance.
Caller's phone number shows incorrectly
Some telephone providers mask the caller number during forwarding, replacing it with the restaurant or business number. Contact your telephone provider and ask them to pass the caller’s real number through instead of overwriting it.
Additional troubleshooting resources
For more specific issues, see:
- Performance monitoring – latency, ASR, and quality issues
- Function maintenance – debugging function errors
- Voice and audio updates – audio quality and pronunciation
- Version management – deployment and rollback issues
- Smart Analyst – analyze patterns across hundreds of conversations to identify root causes
When to escalate
Contact PolyAI support when:
- You see repeated 4XX/5XX errors in call logs that persist after retrying
- SIP handoffs fail consistently and you’ve validated the number, URI, and headers
- Dashboards or transcripts aren’t loading at all
- Call volume drops more than 30% unexpectedly
- The agent’s error rate exceeds 5% across conversations
- You need changes beyond what’s available in the UI

