aws-samples / aws-samples/amazon-transcribe-post-call-analytics

Incorrect Agent/Customer Identification

Open
#339 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
109
Forks
89
PR merge metrics
No merged PRs in 30d

Description

Hi,

In the Post Call Analytics Web UI, we are seeing cases where the customer is identified as the agent and the agent is identified as the customer.
There is no consistent pattern in our call recordings regarding who starts the conversation—sometimes the agent speaks first and sometimes the customer does. The Web UI provides a “Swap Agent/Caller” option at the top right, which manually corrects the labels.
However, we want the speaker identification to be correct by default, without requiring manual swapping.

Is there any way to automatically and reliably identify speakers as Agent or Customer by default in Post Call Analytics, regardless of who starts the conversation?

Contributor guide

Open the contributing guide

Research direction

No file or test is named. Start by tracing the Post Call Analytics Web UI’s “Swap Agent/Caller” behavior and the underlying speaker-label flow, then determine whether Agent/Customer assignment can be reliable when either party starts. Done means documenting a supported automatic approach or the constraints that prevent one.

Written by the indexing model from the issue text.

Assessment

Tech stack
aws, python
Domain
backend-api-design
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.