How to build an agent from recorded audio in VocalLabs

Step by step

1. 1

   Then, upload all your audio files at once. We'll just upload one for now. After that, just click 'Start Processing' to begin the creation. It'll first upload the audio to storage, then generate the transcript. This step identifies our agent and the user in the conversation.

2. 2

   And finally, it builds your agent. This whole process takes a little time to complete, but soon your agent will be ready. Your agent will be generated. As you can see here, the full transcript of the audio you uploaded is now successfully generated and displayed.

3. 3

   It automatically identifies both the customer and the agent within the transcript. You can also manually choose and adjust who is the customer and who is the agent. You can also select the language for your agent. After this, you can set the agent's name, a welcome message, a callback message, and other language settings.

4. 4

   For example, if you want to use Hindi, you simply choose Hindi from the dropdown menu. This supports multilingual agent capabilities. You can also generate the call flow to see how the conversation looks.

5. 5

   Generating the entire call flow takes a moment, as the system processes the interactions to visualize the complete conversational path. While the system processes the call flow, you can explore the various options and settings available for your agent. This includes reviewing prompts and other configurable behaviors to ensure it performs optimally.

6. 6

   That's it! If you click on 'Create Agent', the system will finalize and create your agent. It's a really easy and straightforward process.