### Answer To prevent your agent from interrupting users during ID collection, adjust the interruption_sensitivity setting. ## Lower Interruption Sensitivity This setting controls how sensitive the agent is to user interruptions. [(1) ]() [(2) ]() - Value ranges from [0,1]. [(1) ]() - Lower value means it will take longer/more words for the user to interrupt the agent, while higher value means it's easier for the user to interrupt. - When set to 0, the agent would never be interrupted. For your use case (ID collection with natural pauses) , try lowering this value (e.g., 0.3–0.5) so the agent waits longer before assuming the user is done speaking. ## Adjust Responsiveness You can also tune the responsiveness setting: [(1) ]() [(2) ]() - Value ranges from [0,1]. [(1) ]() - Lower value means a less responsive agent (waits more, responds slower) , while higher value means faster exchanges. Lowering responsiveness gives users more time to continue speaking after brief pauses. ## Context on Turn-Taking Retell AI's turn-taking model analyzes conversational cues (pitch, intonation, timing) to determine when a speaker has finished. [(3) ]() It waits patiently if the user hasn't finished talking, reducing the risk of cutting them off mid-sentence. However, fine-tuning the above settings helps optimize for specific scenarios like digit collection.

​ To learn more, see the sources I considered: [[1] List Voice Agents](https://docs.retellai.com/api-references/list-agents) [[2] Get Voice Agent](https://docs.retellai.com/api-references/get-agent) [[3] VAD vs. Turn-Taking Endpoints in Conversational AI | Retell AI](https://www.retellai.com/blog/vad-vs-turn-taking-end-point-in-conversational-ai) ​

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

### Answer With interruption_sensitivity at 0.65 and responsiveness at 0.95, your agent is configured to respond very quickly (high responsiveness) while still being moderately sensitive to interruptions. [(1) ]() ## Recommended Adjustments For ID collection scenarios with natural pauses, try: - Lower responsiveness (e.g., 0.5–0.7) : Reducing responsiveness by 0.1 adds approximately 0.5 seconds of agent wait time. [(2) ]() This gives users more time to continue speaking after brief pauses like "Yes... [pause]... my ID is..." - Lower interruption_sensitivity further (e.g., 0.3–0.5) : A lower value means it takes longer/more words for the agent to detect an interruption, so it won't jump in during short pauses. [(1) ]() [(3) ]() ## Why This Helps Retell AI's turn-taking model analyzes contextual cues to determine when a speaker has finished. [(4) ]() [(5) ]() It waits patiently if the user hasn't finished talking, but with high responsiveness (0.95) , the agent is optimized for fast exchanges rather than patient waiting. [(1) ]() Lowering both values should significantly reduce premature interruptions during digit collection.

​ To learn more, see the sources I considered: [[1] List Voice Agents](https://docs.retellai.com/api-references/list-agents) [[2] Step 2: Configure the basic settings](https://docs.retellai.com/build/single-multi-prompt/configure-basic-settings) [[3] Update Voice Agent](https://docs.retellai.com/api-references/update-agent) [[4] Retell AI's Turn-Taking Model for Seamless Calls | Retell AI](https://www.retellai.com/blog/how-retell-ais-turn-taking-model-ensures-seamless-calls) [[5] VAD vs. Turn-Taking Endpoints in Conversational AI | Retell AI](https://www.retellai.com/blog/vad-vs-turn-taking-end-point-in-conversational-ai) ​

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