5x NEW VOICES just released: th-cam.com/video/PTCpw1Y9HOQ/w-d-xo.htmlsi=roHjjllMKNHNzLGu 🔗 Make.com BLUEPRINT: bartslodyczka.gumroad.com/l/gmmkvc 🤝 Sign up to Make.com: www.make.com/en/register?pc=bartslodyczka 📺 AI SMS Assistant: th-cam.com/video/HYPw8TfL2Pg/w-d-xo.htmlsi=CVAzhuQzsXH5T2Wa 📋 Take This Quick Survey: forms.gle/otAr1xUamgyYZE5y7
I haven't deployed this yet, but I am working on a different Retell AI tutorial video. A general one about how to make a caller for (I think I will do) a carpet/ rug cleaner business. Should be out in a week :)
I'd love to see this project evolve to do outbound as well with the same simple approach you've taken to the other parts in this series! Also I'd love to see functionality to transfer a call to a live agent or external phone number.
Might be building out something for a client soon that uses Google calendar - but might be using bland or vapi for that. if it's easy enough to transger across to realtime then I'll make it too!
I have one more point to address. Currently, the main drawback of the Realtime API with Twilio is the difference in audio quality when transitioning from the phone (low quality) to the Realtime API. To improve the accuracy of the Realtime API’s understanding, do you have any advice or suggestions on enhancing communication quality? Often, when I say something, the API misinterprets it. This is a significant issue for production scenarios. Could you suggest ways to improve the code? For instance, implementing feedback from the STT flow on top of the Realtime API to ensure understanding, or applying a noise-canceling algorithm to the G711_ULAW audio file.
Great Educational content, thank you Bart. I would love to see the outbound agent as well as the CRM, great ideas on expanding different use cases with this is interesting and thought provoking.
Great video and thanks for the detail walkthrough! I am trying to add additional data points to store in google sheets eg. data3 and data4. How would i add these variables in? I see from your video that the data1 is a solid red color whereas mine is outlined in red. I have a Pro plan already
Hey Kevin, to add more variables you would make sure to add data3 and data4 to every API call within Replit so that they are all consistent. Then, go to your make.com scenario and go to your webhook module, you will need to redetermine the webhook data structure (basically, you need to send an API call that now has all 4 data points so that the webhook updates to expect this new payload). Hope this makes sense? let me know how you go!
Again a great video Bart! Thank you for making it. I'd love to see Part-4 to expand the same prototype with an outbound call from the AI Agent. Also, would it be cool to wrap this into a UI interface? and user-can interact with the AI Agent via chat like interface.
5x NEW VOICES just released: th-cam.com/video/PTCpw1Y9HOQ/w-d-xo.htmlsi=roHjjllMKNHNzLGu
🔗 Make.com BLUEPRINT: bartslodyczka.gumroad.com/l/gmmkvc
🤝 Sign up to Make.com: www.make.com/en/register?pc=bartslodyczka
📺 AI SMS Assistant: th-cam.com/video/HYPw8TfL2Pg/w-d-xo.htmlsi=CVAzhuQzsXH5T2Wa
📋 Take This Quick Survey: forms.gle/otAr1xUamgyYZE5y7
Bought it! Thank you for putting it together!
I would like to see the outbound call example next. Thanks
Will pencil this in 💪
Legend, I had a question regarding the minute 12:51 of your video. Where is the "bart-retell-demo" assistant is deployed? Thanks Bart!
I haven't deployed this yet, but I am working on a different Retell AI tutorial video. A general one about how to make a caller for (I think I will do) a carpet/ rug cleaner business. Should be out in a week :)
I'd love to see this project evolve to do outbound as well with the same simple approach you've taken to the other parts in this series! Also I'd love to see functionality to transfer a call to a live agent or external phone number.
Great suggestion with live call transfer! Will look into this and pencil in a part 4 :)
Hey Bart any chance you can do one with Google Calendar to create new appointments/change/cancel? Thanks heaps
Might be building out something for a client soon that uses Google calendar - but might be using bland or vapi for that. if it's easy enough to transger across to realtime then I'll make it too!
@@BartSlodyczka Thanks Bart just took your survey
I have one more point to address. Currently, the main drawback of the Realtime API with Twilio is the difference in audio quality when transitioning from the phone (low quality) to the Realtime API. To improve the accuracy of the Realtime API’s understanding, do you have any advice or suggestions on enhancing communication quality? Often, when I say something, the API misinterprets it. This is a significant issue for production scenarios. Could you suggest ways to improve the code? For instance, implementing feedback from the STT flow on top of the Realtime API to ensure understanding, or applying a noise-canceling algorithm to the G711_ULAW audio file.
can you also share the Retail AI information you mentioned?
Will look into this! 🙏
Great Educational content, thank you Bart. I would love to see the outbound agent as well as the CRM, great ideas on expanding different use cases with this is interesting and thought provoking.
Thank you very much my man! Will work on something soon, keep killing it 💪
another awesome video! thanks man!
Thanks legend!!
Hi, thank you for video, very nice and detailed explanation, patiently waiting for OpenAI API Masterclass: Platform, Models & API - Part2 - Part5 👍
Thank you very much :) Working on Part 2 and will release shortly 💪
proper wild tutorial Vi-Lo
Thank you boiz 💪
Great video and thanks for the detail walkthrough! I am trying to add additional data points to store in google sheets eg. data3 and data4. How would i add these variables in? I see from your video that the data1 is a solid red color whereas mine is outlined in red. I have a Pro plan already
Hey Kevin, to add more variables you would make sure to add data3 and data4 to every API call within Replit so that they are all consistent. Then, go to your make.com scenario and go to your webhook module, you will need to redetermine the webhook data structure (basically, you need to send an API call that now has all 4 data points so that the webhook updates to expect this new payload). Hope this makes sense? let me know how you go!
Again a great video Bart! Thank you for making it.
I'd love to see Part-4 to expand the same prototype with an outbound call from the AI Agent. Also, would it be cool to wrap this into a UI interface? and user-can interact with the AI Agent via chat like interface.
Thanks my man! Great suggestion via wrapped UI, I’ll do some thinking on this :)
@@BartSlodyczka yes that would be awesome ! Maybe by using replit AI assistant to build the UI and connect the dot
♥