### Answer Yes, there can be differences between web/test calls and real phone calls. Phone calls involve additional telephony infrastructure that web calls don't—this includes network conditions, carrier routing, and audio quality variations that can affect how the turn-taking model interprets speech and silence. [(1) ]() Retell AI uses a turn-taking model that combines Voice Activity Detection (VAD) with contextual analysis to determine when a user has finished speaking. On phone calls, factors like background noise, line quality, or brief silences may be interpreted differently than in a controlled web call environment, potentially causing the agent to misinterpret pauses as the end of a conversation. A few things to check: - Review the

disconnection_reason

in your call data—this field will tell you exactly why the call ended (e.g.,

agent_hangup

,

user_hangup

, or

error

) . [(2) ]() - Check the public call logs via the

public_log_url

to see the detailed request/response flow and latency tracking, which can help identify where issues occur. - **Consider the turn-taking model's behavior**—it predicts when to take turns and may respond differently to real phone audio versus web audio. The model learns and adapts to different conversational styles. [(3) ]() If the agent is hanging up unexpectedly, the call logs should reveal whether it's a turn-taking misinterpretation, a tool call issue, or something else in the conversation flow.

​ To learn more, see the sources I considered: [[1] VAD vs. Turn-Taking Endpoints in Conversational AI | Retell AI](https://www.retellai.com/blog/vad-vs-turn-taking-end-point-in-conversational-ai) [[2] Register Phone Call](https://docs.retellai.com/api-references/register-phone-call) [[3] Performance-related configurations and noticeable latency during live calls](https://community.retellai.com/t/performance-related-configurations-and-noticeable-latency-during-live-calls/747) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.