- Newest
- Most votes
- Most comments
Note that Contact lens starts working ONLY after the customer's call gets connected to a live agent. You might be able to achieve dynamic routing through a different mechanism like Lex, Customer profile, Case or your own database lookup.
Also, once connected to agent, it's possible to stream and display real-time transcript/categories detected if you are using a custom CCP. Refer https://aws.amazon.com/blogs/contact-center/contact-lens-real-time/
So let’s start with where would you get the insights from? I assume you mean Lex or are you trying to get the insights from when the caller is talking to the agent as that seems to defeat the purpose. If you are getting the insights from Lex, you already have the intents you created so why not map those out to routing profiles ahead of time and avoid having to build this process?
David
Thank you for your response and for seeking clarification. You've brought up an important point.
In our case, we are indeed exploring the possibility of obtaining insights directly from the conversation between the caller and the agent in real-time using Amazon Connect Contact Lens. We understand that Amazon Lex provides predefined intents, but our goal is to analyze the actual dialogue during calls to identify specific keywords, phrases, or sentiment indicators that may not be covered by predefined intents
Using Lex V2 you can find out customer sentiment and then route call based on customer sentiment. Refer https://docs.aws.amazon.com/lexv2/latest/dg/sentiment.html
Relevant content
- asked 3 months ago
- asked 2 years ago
AWS OFFICIALUpdated a year ago- AWS OFFICIALUpdated 5 months ago
- AWS OFFICIALUpdated 2 years ago
- AWS OFFICIALUpdated a year ago
Thank you for your response and the valuable information you provided regarding Amazon Connect Contact Lens.
To clarify further, is it possible to route contacts dynamically based on customer sentiment detected by Contact Lens? In our scenario, we are interested in using sentiment indicators (positive/negative sentiment) to influence the routing of contacts to specific agent groups or handling processes.
As discussed, we're keen on leveraging Contact Lens for call analytics, particularly its ability to identify keywords, phrases, and sentiment during live agent-customer interactions. If routing based on sentiment isn't a direct feature of Contact Lens, we'd appreciate any insights on alternative methods or best practices to achieve this sentiment-based routing within the Amazon Connect ecosystem.
Not that the user posted that contact lens starts working after the call already landed with the agent and if I remember correctly it's recommended for the transcription to run after the call ends for better accuracy, so what you're trying to do would not work exactly as you expect as the caller is already talking to an agent. But maybe you're trying to do something like this (https://docs.aws.amazon.com/connect/latest/adminguide/contact-lens-rules-email.html) based on a rule you could then create some sort of escalation/barge in mechanism to triage a call that is not going well.
david