### Answer When STT (speech-to-text) produces garbled results for addresses, here are several approaches you can use: 1. Use Boosted Keywords Add address-related vocabulary (street names, city names, common address terms) as boosted keywords to bias the ASR toward domain-specific terms for higher transcription accuracy. [(1) ]() 2. Enable Advanced Denoising Mode You can choose between two levels of noise cancellation: - **Remove noise**: Filters out background sounds - **Remove noise + background speech**: Filters out both noise and unwanted speech This helps boost transcription accuracy in noisy environments. 3. Implement Fallback Strategies Use prompt engineering techniques to have the agent ask follow-up questions when uncertainty arises. When the AI is unsure of a user's meaning, it should ask clarifying questions to double-check and prevent misunderstandings. [(2) ]() *4. Use User Keypad Input (DTMF) * For critical data like addresses, consider allowing users to spell out or confirm via keypad input. You can configure: [(1) ]() - End with a special key (# or *) - Fixed digit length - Timeout-based ending *5. Enable Medical Transcription Mode (if applicable) * If you're in healthcare, there's a medical vocabulary setting that improves accuracy for specialized terminology. **Practical tip**: Have your agent repeat back the address and ask for confirmation, or request the user spell out difficult parts (street names, apartment numbers) . Check out these potentially useful sources: - [Troubleshooting AI Voice Agent Issues]() - [Platform Changelogs]() Would you like

a general answer on additional STT improvement techniques?

​ To learn more, see the sources I considered: [[1] Platform Changelogs | Retell AI](https://www.retellai.com/changelog) [[2] Troubleshooting AI Voice Agent Issues | Retell AI](https://www.retellai.com/blog/troubleshooting-common-issues-in-voice-agent-development) ​

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