Skip to main content

​
Feature Release

This is the announcement for an upcoming ASAPP feature. Your ASAPP account team will provide a target release date and can direct you to more detailed information as needed.

​
Overview

ASAPP is adding an AI Transcribe implementation pattern for Amazon Connect. ASAPP’s Amazon Connect Media Gateway will allow Kinesis Video Streams audio to be easily sent to AI Transcribe. A similar call signaling integration via the Start/Stop API will be leveraged as is used for the integration with the SIPREC Media Gateway.
Amazon Connect Lambda functions send requests to start and stop transcription. ASAPP requests streaming media from Kinesis Video Streams and produces transcripts for real-time, after-call, and batch use cases.

​
Use and Impact

The new Media Gateway will allow for a simplified integration for customers leveraging Amazon Connect as their CCaaS provider, reducing time and effort of sending call media to ASAPP.

​
How It Works

Procedure of streaming audio to ASAPP
  1. Setup IAM roles for authentication with Kinesis Video Streams and enable media streaming.
  2. Configure triggers to start and stop media streaming in Connect flows.
  3. Send requests to start and stop transcription using AWS Lambda functions, including parameters that ASAPP uses to request media from Kinesis.
  4. ASAPP requests media from Kinesis Video Streams and begins transcribing upon acquiring media.
  5. Receive transcript outputs leveraging one of ASAPP’s transcription delivery mechanisms.
Transcription Settings Transcription settings (e.g., redaction, language) are configured as part of requests to the /start-streaming endpoint and will be reflected in the messages returned from ASAPP. No further configuration is required.

​
FAQs

Is integration with the Start/Stop API required? Yes. Start requests to the API provide ASAPP with required call attributes to request customer and agent audio from Kinesis Video Streams. Stop requests are also required, as they allow AI Transcribe to exclude hold and queue audio from transcription. This is a requirement for downstream services to function properly, ensuring transcription is only leveraged where necessary.