Log out agents automatically when they close their CCP

When using the default Amazon Connect CCP, closing the CCP window doesn't change an agent's status from Available to Offline. An agent must change their status manually, and then log out.

To change this behavior, you must create a custom CCP.

Use the Amazon Connect Streams API and the Agent API to create a custom CCP for your contact center. For an example custom CCP setup, see Perform an external screen pop with Amazon Connect.

Use the following steps to update your CCP so it logs out agents automatically when they close the CCP window.

Step 1: Set up the Streams API

For instructions, see the Amazon Connect Streams Documentation.

Step 2: Update your application code to change the agent state

Integrate the following Streams API calls into your web application:

1. Use connect.agent() to subscribe to agent events and retrieve agent objects.

   let mAgent;
   
   connect.agent(function(agent) {
     mAgent = agent;
   });  
                           
                       

2. Call agent.setState() in the onbeforeunload event handler to change the agent state.

   window.addEventListener("beforeunload", function(event) {
       if (mAgent != null) {
           let states = mAgent.getAgentStates();
           // "states" is an array of changeable states. You can filter the desired state to change by name.
           let offlineState = states.filter(state => state.name === "Offline")[0];
                         
           // Change agent state
           mAgent.setState(offlineState, {
           success: function() {
               console.log("SetState succeeded");
           },
           failure: function() {
           console.log("SetState failed");
           }
       });
     }
   });
                           
                       

Step 3: Design for errors

If an API call fails to execute the first time and a contact takes the error branch of your contact flow, there's a chance that an agent's state won't change as expected. Be sure to include logic to account for this possibility. For example, you could delay the page unload while the API call is tried again. Or, you could pop a "Call failed" warning message in a modal dialog before the page unload.