Configuring and running an Agent Trainer simulation

This demonstration shows you how to configure scenarios and run a simulation for an Agent Trainer.

U+ Bank has added the Pega GenAI Agent Trainer™ connection to train agents in handling common customer service interactions such as address changes, filing complaints, and account closures through its Digital Messaging Channel interface.

Now, the bank wants to configure realistic scenarios for the Agent Trainer and run a simulation to ensure that their agents can practice and refine their skills in a realistic, risk-free environment.

To configure Agent Trainer settings, on the Connection tab of the Digital Messaging Channel interface, click Manage connections. 

In the Digital Messaging Manager window, click your Agent Trainer connection to configure the Agent Trainer simulation options.

Enter the name of the organization for the Agent Trainer. The system uses the organization name so that the simulated customers can form business-specific contexts around the tasks that the system assigns to them during a simulation.

In the Escalation Command field, enter the escalation command to simulate in your Agent Trainer. The escalation command must correspond to an existing create case command that you configured in your Digital Messaging channel. In this example, enter escalate as an escalation command.

Specify the number of sample customers for the simulation. When you begin a simulation, the system generates the specified number of customers in the Agent Trainer.

Now configure simulation scenarios for your Agent Trainer to run a simulation for the Agent Trainer.

Because the bank wants to train agents in handling common customer service interactions such as address changes, filing complaints, and account closures through its Digital Messaging Channel interface, add simulation scenarios for all of them.

First, add a simulation scenario to update an address. In the Simulation Scenario, enter a clear description of a scenario for the simulation. Be clear about the scenario or task that the system assigns to the simulated customer to complete during their interaction.

In the Queue Selection field, enter the queue selection for the scenario. You identify the specific Pega queue that the simulated customer selects for this scenario during their interaction.

In the Personality field, enter a description of a personality for a customer. Consider attributes such as patience, urgency, or technology-savviness, which guide the tone and interaction style of the simulation.

In the Pain Points field, enter the pain points information, which reflect the challenges or frustrations that customers might have. Pain points might include difficulties with navigating the system, concerns about data accuracy, or time constraints. These modifiers help the prompt to better simulate a realistic customer experience.

Continue to add simulation scenarios to file a complaint and close an account.

After you finish configuring the Agent Trainer, it is time to run a simulation.

Before you run the simulation, log into the Interaction Portal as a CSR to test the scenarios.

Return to the Digital Messaging Manager portal, click Start simulation, and then click the Live simulation tab. The system displays the real-time simulation taking place between AI-generated customers and real-life CSRs. It also shows the AI-generated customer name and their status, as well as additional data related to their simulated task. All interactions are stored in the Pega Customer Service™ application.

As CSRs engage in this simulated chat interaction, Pega Customer Service handles it just like any other customer chat, which enables the manager to deliver a real-time experience. This setup also facilitates the interaction analysis and the provision of real-time feedback.

Return to the Interaction Portal to serve the request. The CSR receives an incoming call from the AI-generated customer. The CSR accepts the call, and then begins the interaction with the customer, and solves the request.

After the end of the conversation, the entire transcript goes to the feedback engine, where you can generate metrics and identify strengths and weaknesses for the CSR based on each interaction.

While the performance evaluation might not always be flawless, the feedback engine offers valuable insight through various scores, such as an assistant score, a professionalism and efficiency score, and a customer satisfaction score. It also highlights some strengths and weaknesses, all of which are recorded and saved in the simulation for further reporting in Pega Customer Service.

You can view the current AI-generated simulation data by clicking the Simulation history tab. You can download the file to view the data points relevant to the simulation.

You have reached the end of this video. You have learned:

• How to configure scenarios for the Agent Trainer.
• How to run a simulation to train your CSRs in handling customer interactions.