- Newest
- Most votes
- Most comments
The contact object is only available if you use Amazon Connect Streams API. If all you want to do is display contact information in a view as the call is presented to the agent you do not need 3P apps. You can create a simple Step-by-step guide view and display the information you need as soon as the call connects.
If you need a foundational understanding of the step by step guides check out this course: https://explore.skillbuilder.aws/learn/course/internal/view/elearning/20251/amazon-connect-flow-modules-and-step-by-step-guides.
For practical exercises check out this workshop: https://catalog.workshops.aws/amazon-connect-agent-empowerment/en-US. The workshop shows you how to create step-by-step guides, but also how to integrate a 3P application in the agent workspace.
And another helpful blogpost. https://aws.amazon.com/blogs/contact-center/getting-started-with-third-party-applications-in-the-agent-workspace/
If you provide more information about your specific use case, we can help you narrow down your best approach for the challenge you are trying to solve.
Amazon Connect does not support OpenID authentication, so Streams integration using hidden CCP for event handling is not your best integration option here.
You can use the events available for the 3P application integrations: https://docs.aws.amazon.com/agentworkspace/latest/devguide/integrate-with-agent-data.html. Also, check out this documentation page that describes the events available: https://docs.aws.amazon.com/connect/latest/adminguide/3p-apps-events-requests.html.
I am not sure what framework you use for your application, so I cannot tell if this ^^ integration method works for you, but it may work. Cover all the information in the Getting started with third-party development in the Amazon Connect Agent Workspace: https://docs.aws.amazon.com/agentworkspace/latest/devguide/getting-started.html to better understand how applications are loaded in the agent workspace, as well as 3P APIs and requests that give you access to the contact data.
If your current application architecture does not allow this level of integration (make sure your app is iFrame-able), you can always revert to a data backplane where you store the current contact data in a backend DB and you restore it as the call is presented to the desktop app using whatever means you have available in your app. The trick here will be to find the unique identifier. Depending on the timing of your data display (ringing or connect) you can use as key for the data the agent id that gets the incoming contact. In the agent whisper flow you can save the contact Id with all the contact info and the id of the agent that will be available in the Agent property of the whisper flow. You app will need to query by agent Id, so that means you will need to have a "translation" of your agent Id unless the OpenID and Connect user use different IDs for the agent.
I hope this helps.
Relevant content
- asked a year ago

Thanks. I have an existing real-time application (potential 3rd party app) that uses business logic to construct a data set around the CALLER. It needs the caller profile details such as the caller id/phone number, the time the call came in, and the Agent details of the agent who answered the call. Additionally that app also uses SignalR to display some dynamic information and has an internal communication tool.
Instead of folding everything inside a VIEW in Agent Work Space which could require redoing the app, the hope was to launch the app inside the Agent Work space which as we understand would runs inside an iframe. I was asking if there were some well known functions one could call on the parent window to get the kind of details we would need to personalize our 3rd party app.
If we have to use connect streams, but dont need to load the CCP widget in our app:
does that cause performance or memory load on the 3rd party app ?
Are any media streams for the call in progress sent to the 3rd party app even though not consumed?
Our app uses OpenID connect for its own authentication, which is different from the auth used by the Amazon Connect as it is run by a different party. So Agent can potentially have two auth flows (another reason to not load the CCP inside our app if it can be avoided)
Hope that clarifies the ask?