Here’s the video: https://youtu.be/ZaXhBWnp-Y8?si=S-HtVgBihocuzb9L Maybe I'm late to the game on this one, but got Assistable replying to emails and it's awesome. Here's my specific use case. Inbound email --> GPT determines if it's a general inquiry (AI responds) or about a specific project. If general inquiry, Assistable replies. If someone needs to reply, it goes down Human Response path and the client's team gets notified. Let's gooooooooooooooo

The Get Conversation premium tool doesn't work and it would be awesome to have! We could feed SMS history into our prompt so Voice doesn't need to go over fields that are already covered. I believe it's a GHL issue. I tried recreating with webhooks and always end up with some errors. Usually, an Invalid JWT error. @Jorden Williams can you take a look at this and see if it's something you can resolve? Would be an incredibly big boost to the AI logic. Thank you!

So lately, I have been noticing the AI will go silent during the call for 3-5 seconds. This happens with and without a set task. Is anyone else experiencing this? Does anyone have a solution for this yet? This is irritating a lot of clients.

We build a simple self schedule to call the contact back with one sentence of prompt. This video goes over a couple of things: 1. How our decoupled tooling allows for cool solutions 2. Why to keep it simple and how we make it easy to do that. 3. How to better understand your AI for production cases 4. Take a look at how to read tool call meta data to better prompt your AI, and under stand the thought process of your AI through our internal monolog and context system.

- Location ID based rate limiting system - Get conversation logs - Make AI Call in draft to be submitted for release - URL translation and summary action - Run bot action So, everything is staged - there are couple of datathings that we are putting together and then hoping for final push tonight which should seperate operations from the front end completely Our inbound script got a lot faster so we should expect to see 100% pickup rates with fallbacks on inbound We also reformatted the 'Make AI Call' action with the new outbound script that does the same thing - we even will send back data to GHL to let you know the call id, status, success status, error messages and returned an error which you can use to build if elses - all will be released in a doc and a basic snapshot you can use to set up a calling flow Chat runs are completely staged in migration which allow for contact or account based timezone booking, faster processing & 3 LLM status checks for fallbacks so we can have (theorically) no downtime for a chat run same for voice, we will have a fallback calendar ID, fallback voices, fallback LLMs and fallback webhooks, no-access token staging, notifications for oauth disconnection and outages - so that there can be handling as much as we can possibly do Working on GHL actions for calling based services for client reporting & working on a call finished trigger so you dont need a post call webhook as well Will update once testing has concluded