Role Play (Overview & Set up)

Overview:

The Role Play workflow in Bongo lets users practice real-world conversational skills through interactive AI-powered simulations. This assessment type enables users to engage in dynamic, two-way conversations, where they must adapt their communication in response to real-time feedback from an AI agent.

In the Role Play Workflow:

• Users engage in live conversations with an AI agent that represents a realistic persona or stakeholder.  They can practice as many times as they'd like and receive feedback and coaching after each interaction.
• The AI agent follows predefined behaviors, concerns, and personality traits set by the assignment builder
• Conversations are evaluated using AI feedback, human evaluation, or both

Access requires activation and may involve additional licensing fees. Please contact Bongo for more information: [e-mail]

Set up & Configuration:

Configuration for Role Play assignments is split up into tabs:

The General tab contains basic information about the assignment

1. Your LMS may populate the assignment title, but otherwise, you will need to add it.
2. You can optionally add a due date to the assignment. The due date is informational for users and does not prevent submissions after it passes. Evaluations may choose to adjust scores for late submissions at their discretion.
3. Instructions are the first thing the user will see when taking the assessment. Use this space to provide any necessary context or guidance for users before they start the assessment. A video can be included with the instructions by clicking on the video icon next to the instructions. [DO NOT use emojis in your instructions].

More Options

In More Options, you can create Post-Submission Instructions. This is an optional message that users will see after completing the assessment. There are also options to configure when certain information is available to users.

The Evaluation tab determines how users will be evaluated. Role Play offers Smart Scoring or Auto-Pass options. Smart Scoring is the recommended evaluation method when you want to automate scoring and tie the feedback to your defined Learning Objectives and Criteria.

Evaluation Type Considerations:

Smart Scoring

• Smart Scoring uses the criteria defined in the Learning Objectives tab to automatically score a user’s submission.
• Evaluators can still determine the final score, review responses, and provide comments.

The Role Play tab is where you define the interactive elements that drive the AI agent's behavior during the conversation.

1. In the The User/Learner Role field, clearly define who the user is (e.g., "You are a Channel Account Manager at a technology reseller") and the objective of the conversation (e.g., "Your goal is to position your solution's value and address concerns about cost and implementation time")
   
   Note: Users do not see this information. Any instructions that the user should see should be included in the Assignment Instructions in the General tab.

2. In the AI Agent Role field, define the persona and context for the AI agent in the scenario. This establishes who the AI will portray during the conversation, their background, and their current situation.
    
3. In the AI Agent Key Concerns field, provide a list of concerns, objections, or priorities that the AI agent should consider during the interaction.
    
4. The Conversation Guidelines field defines the rules and structure of the scenario to guide the conversation's flow.
    
5. Maximum Duration defines how long the session is allowed to go. If this time nears during a session, the AI agent will include a brief warning that they don’t have much time left in their responses to the user, and if the maximum time is passed, the AI agent will let the user finish talking, but will then say goodbye and end the session.
    
6. Select the voice for the AI Agent from the available options. Descriptions and voice samples are provided for each voice. The image associated with the selected voice is also what users will see during each session.

Within the Learning Objectives tab, you can generate a list of Learning Objectives.

Learning objectives define what users are expected to say or how they are expected to react during the role play session. Each learning objective includes a general description and supporting criteria that outline the key concepts, actions, or behaviors that a strong response should include.

When learning objectives are paired with Smart Scoring, the supporting criteria are also used to automatically evaluate submissions. This allows evaluators to reduce manual scoring while maintaining consistent standards.

Learning objectives can be paired with the Auto-Pass evaluation type to provide more targeted feedback, while not using the objectives for the final evaluation.

To add Learning Objectives, you can provide reference material, and Bongo will automatically generate a list of objectives and criteria based on the material provided.

The generated objectives can then be modified or removed if desired. Either deselect any generated criteria that you do not wish to use, or once generated, you can modify them by hovering your cursor beside the criteria, and you will see an edit icon.