Event bridge has a clear topic for voice transcription monitoring. But what topic would you recommend to get real-time events of Webmessaging, or SMS, or digital events?
Thanks
Transcriptions are voice-only because every other media format is already in text format and doesn't need transcription. Are you looking for the recordings for a conversation? For that, you would use an appropriate conversation event for your situation and once it's over, start using the Recording APIs to fetch the recordings the normal way. You can use the topic v2.users.{id}.conversations.{id}.recordings.{id} to know when transcoding of the recording is complete and the file is ready for download if you prefer not to poll, though I think this event is only relevant to voice because I believe only audio formats require transcoding.
Hi Tim,
I'm looking to get real-time communication for an active digital channel conversation. For example, as an agent is working an SMS I would like to get the text exchanged as close to real-time as possible.
You can only do this if your application is the user that's actively participating in the conversation. You will need to implement the specific media handling for each media type you wish to support in that way. If you're having trouble finding the docs for a specific media type, let me know.
What you cannot do is monitor the messages in a conversation using client credentials or as a user that is not an active participant.
OK thanks can you share the docs for say SMS.
Your user application would subscribe to the v2.users.{id}.conversations.messages websocket topic for the conversation. To use the API to get a message in the conversation, use the conversation metadata to get the parameters for GET /api/v2/conversations/messages/{conversationId}/messages/{messageId}.
Thanks Tim,
I was afraid we were going to need to do a lot of GET's. Everyone exchange would result in a GET multiplied by the number of interactions could potentially create a lot of calls to the API and get us into the FAIR USE world. Unless I'm looking at that wrong... How does agent assist get around that or what API's does it hook into?
Fair use is only measured for your custom applications. Applications built by Genesys do not count towards your quota. You can find the "API on demand" fair use policy here: Genesys Cloud fair use policy - Genesys Cloud Resource Center.
Doing some quick napkin math, a CX2 user gets 110,000 requests/month. Allocate ~10% of that for overhead, so 100k for usage. If you make 10 API requests per message, that's 10,000 messages to hit the limit, or the agent handling a new message every 1.04 minutes 40 hours per week for an entire month. That doesn't seem like a particularly likely scenario to me, but if you do have agents working a lot of overtime or your app is really chatty with the API, it could be a concern. If you do this math with the current workloads for your agents and how many requests you think it would take to accomplish your goal, it should give you a reasonable idea of what kind of usage you're looking at for your application.
Keep in mind that this fair use allocation is pooled for the entire organization and is shared between all of your custom apps, so it's not a per user calculation as done above.
Thanks, Tim. Have a great weekend.
1 Like
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.