V20 Update 8 Alpha - The AI Chapter

Hello everyone!
Will I need a subscription to record calls in stereo mode (just recording and storing)?
Or will it be available to everyone?
 
Having an issue with the transcription, and not sure where to dig and troubleshoot.

AI Agent is working, however I have set to recording, and added the local transcription, added a dedicated GPU server, connection shows as up and available, however no transcriptions coming through. If I change to 3CX cloud, it will transcribe.

Server is a Single Socket; Xeon Gold 6242R (20 core)
192GB RAM
Quadro RTX6000 GPU

SNGOUT2137.jpg

SNGOUT2136.jpg
 
  • Like
Reactions: nikolascx
Hello unnfortunately with pro edition I cannot test AI receptionist feaure. Is it possible to have temporary Enterprise KEY during alpha/beta period in order to test the new Ai features?
 
  • Like
Reactions: lenny76
Hello everyone!
Will I need a subscription to record calls in stereo mode (just recording and storing)?
Or will it be available to everyone?
Its part of the Enterprise edition....
 
Its part of the Enterprise edition....
I think they meant more like, will the stereo recording be available for other engines like OpenAI/Google? Even some CRMs, like HubSpot, would appreciate the stereo sound to do analysis on their side. For now, it seems to be limited to the 3CX engine.
 
And the synchronization of the department phone book with the IP-phones will come later in Update 8, right?
 
  • Like
Reactions: noord
Hello unnfortunately with pro edition I cannot test AI receptionist feaure. Is it possible to have temporary Enterprise KEY during alpha/beta period in order to test the new Ai features?

have the same idea
want to install a temp system to test it without touching the production system
 
Perhaps a stupid question:
But I have used the AI-agent to leave behind a message. The agent understood (at least, I think it understood) for whom the message was, asked me who I was and my company, asked for the message text and confirmed everything. So that works great!

But: where does the message go??
I expected to get either a chat-message, an email or perhaps a voicemail of some kind. But I didn't get the message anywhere.
Where is it supposed to go, or isn't this part complete yet in the Alpha?
 
Perhaps a stupid question:
But I have used the AI-agent to leave behind a message. The agent understood (at least, I think it understood) for whom the message was, asked me who I was and my company, asked for the message text and confirmed everything. So that works great!

But: where does the message go??
I expected to get either a chat-message, an email or perhaps a voicemail of some kind. But I didn't get the message anywhere.
Where is it supposed to go, or isn't this part complete yet in the Alpha?
Hi @noord , I wonder where are you routing the call? if it goes to an user extension, then he will be screened with that info in the web client when receiving the call.
 
Hi @noord , I wonder where are you routing the call? if it goes to an user extension, then he will be screened with that info in the web client when receiving the call.
No, I just tell the AI-agent that I want to leave a message for <user name>. It then takes the message, confirms it and says it will deliver it. But obviously, it doesn't do that
:)
So, it's not really a transfer, it just when you ask the AI to leave someone a message.
 
After spending some time with the family (Fall-holidays) I came back to this update.
  • Stereo transcription instead of a plain text transcription is awesome.
  • AI Receptionist, 10 seconds of configuration, minutes of fun playing arround with it, and working flawless.
  • Deparment phonebooks awesome to see these coming in, looking forward for more updates and features for these.
Awesome work 3CX team!
 
Hello, Regarding Department Phonebook I have only this options, what goes wrong? (16SC pro)

1762109109008.png
 
I think they meant more like, will the stereo recording be available for other engines like OpenAI/Google? Even some CRMs, like HubSpot, would appreciate the stereo sound to do analysis on their side. For now, it seems to be limited to the 3CX engine.
For now its 3Cx only because there is a time intensive process that gets done after the recording ends and it gets done on the Transcription node or on the cloud service for the moment. As its too disruptive for the PBX core. Also both OpenAI and Google are not able to analyze stereo. I don't see a need to let OpenAI or Google do analysis on the stereo - we as the PBX do this the best with the info we have on the voice streams but you can of course analyze the transcribed diarized text in Google and OpenAI and of course you can then do much more. It may change in the future as we make the process more real time friendly.

@ArneDery - Thank you for the heads-up!
 
Last edited:
For those that are trying out the AI/Transcription with your own hardware, im curious what are your hardware specs and what kind of performance are you getting?

For 3CX staff, will there be a table of hardware compatibility/approx performance expected so we can judge what we should be buying for our specific situation?

The two options mentioned in the two blog posts so far are consumer/Dev desktop hardware (ASUS Ascent GX10) and (5090 GPU).
I'm looking for compatible GPU options suitable for an server.
 
  • Like
Reactions: nikolascx
I have a question regarding hardware compatibility for the 3CX Local Transcribing Server.

In China, NVIDIA is replacing the standard 5090 GPU with the 5090D version for the market.

My question is: Will the NVIDIA 5090D GPU be suitable/capable of running the 3CX Local Transcribing Server?
 
  • Love
Reactions: nikolascx
Hello everyone!
Will I need a subscription to record calls in stereo mode (just recording and storing)?
Or will it be available to everyone?
Hi!
The Stereo transcription only works when you are using 3CX Transcription service.
So you have the option

1762152086220.png
 
I think they meant more like, will the stereo recording be available for other engines like OpenAI/Google? Even some CRMs, like HubSpot, would appreciate the stereo sound to do analysis on their side. For now, it seems to be limited to the 3CX engine.

Hello Team
When we sent a stereo file to Google and openAI they frowned.. They wanted only mono.
But I see openAI now has more models that diarize automatically. So maybe they wont grumble anymore.
At the moment, stereo format is only available with 3CX Transcriber service
 
I have a question regarding hardware compatibility for the 3CX Local Transcribing Server.

In China, NVIDIA is replacing the standard 5090 GPU with the 5090D version for the market.

My question is: Will the NVIDIA 5090D GPU be suitable/capable of running the 3CX Local Transcribing Server?

Hi Dears
We don't know yet. And we cannot know because that chip is specially made to comply with Trumps US export restrictions on high performance AI Chips to China.

If I correctly understood, it looks that Jensen complained to Trump telling him that the ban on China is hurting Nvidia - an American company. So then Trump allowed him to create a chip for China which is inferior to the 5090 but still a 5090 - called the D dragon series.
So this chip is lower in power, but complies with US Regulation export laws.

Now will China accept this chip?
Can you get your hands on one of them right now? If yes, make sure you get the 5090D (Original)
A new 5090 D v2 was created with even STRICTER limitations.

We in the dev team will never be able to get our hands on this chip. You will need to try it.

I made you a quick table explaining the differences
I strongly believe it should work..

Keep me posted - Great topic!!

1762160356682.png
 
The thing is, we use our own transcription solution, on our own servers and neural networks. For us, the most important thing is simply to be able to record stereo files in the PBX file system and then transmit them to the server with our neural networks in their pure form.
So, the question is, will we be able to do this after upgrading from the Pro version to the Enterprise version?
 

Forum statistics

Threads
111,990
Messages
590,165
Members
164,928
Latest member
blackacoustic