Are you wanting it to retrieve the voice style or the content of the audio (like a transcript)?

@Esteban Lopez im wondering if it has something to do with "Buy Workflow Pro Plan" on the ghl said? I think they announced that recently? there might be a ghl setting in the way.

@Tyler Frick if the Ivr takes audio input, you can handle this simply by prompting. Example Prompt for General Navigation: """You are about to call a business that may have an IVR system. Listen carefully to the audio prompts and speak the appropriate digit to navigate to the desired department, such as support or billing.""" 2. Specific Instructions When You Know the IVR Path If you know the digits to press beforehand, customize the prompt to save time and ensure accuracy. Example Prompt for Known IVR Paths: """Press digit 1 to reach the support department. Press digit 2 for billing inquiries.""" Test: • Step 1: Save your agent configuration with the prompts. • Step 2: Test the setup by making an outbound call to a real IVR system. • Step 3: Alternatively, you can simulate the IVR system by calling yourself and saying phrases like “Press 1 for support” to see how the AI responds.

@Marcel Lee 2H 29M mark in @Jason Dueck's recording https://www.skool.com/assistable/build-multi-calendar-booking-now?p=b5a25fcc

@Francisco Magistra best approach is to increase your wait time in the assistant settings by a bit. Though I've seen it accidentally queue them separately anyways, most of the time they concat into a single user input (from the AIs perspective).

@Harry Stokes you're on literal fire these days dude. Thanks for sharing. Gorgeous.

@Harry Stokes I've got a 20million token dataset laying around somewhere from a multi-hour ghl documentation Crawl lol blasted it to 32,000 dimensions (>2gb's of RAW TEXT) and brought it back down to clusters in a 512 X 512 feature matrix for vector embeds - topically organized and cleaned. Herded everything into 12 main clusters. Word clouds below are the most prevalent keywords from each.