Here is a video that was inspired by this question: https://youtu.be/I9Tdmm6qAh8

I’ve been experimenting with creating a single-file HTML5/JavaScript interaction designed to be used as a Web Object in Adobe Captivate 13 (Project Charm). I tried a couple of prompts in Google Gemini that might produce what you’re looking for. Visit Gemini.Google.com (if you have a paid account, you might get better results). Click on Tools and turn on canvas. Then enter the following prompt:

Objective: Act as an E-learning Developer and Full-stack Web Expert. Create a single-file HTML5/JavaScript interaction designed to be used as a Web Object in Adobe Captivate 13 (Project Charm). The interaction must simulate a conversation with a retail customer and use the Gemini API to evaluate the effectiveness of student questions.

Technical Specifications:

1. Single File Requirement: All HTML, CSS, and JavaScript must be contained in a single file. index.html file.
2. AI Integration: – Use the gemini-2.5-flash-preview-09-2025 model via fetch to https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-preview-09-2025:generateContent?key=${apiKey}.
   • Implement exponential backoff for the API (5 retries: 1s, 2s, 4s, 8s, 16s).
   • Use a System Prompt to force the AI to return a JSON object containing: isOpenEnded (boolean), customerResponse (string), and feedback (string).
3. LMS/Authoring Bridge: Use the Captivate 13 API window.parent.cpAPIInterface.next(); to advance the slide once the goal is reached.
4. UI/UX Design: – Modern, clean aesthetic with Tailwind-like CSS or high-quality custom styles.
   • Include a progress tracker (e.g., dots or a bar) showing 0/3 successful questions.
   • Provide a “Next Slide” button that only appears after 3 successful open-ended questions.
   • Include loading states/spinners while the AI is “thinking.”

Pedagogical Logic:

• System Prompt for AI: “You are a customer in a store. Evaluate the student’s question. A question is only successful if it is truly open-ended (encourages a detailed response and cannot be answered with ‘Yes’ or ‘No’). If successful, give a realistic in-character response. If unsuccessful, give constructive coaching feedback.”
• Win Condition: The student must successfully ask 3 open-ended questions to “win” the interaction.

Styling Preferences:

• Responsive layout (works in iframe).
• Rounded corners, subtle shadows, and a clear distinction between “Customer Speech” and “Student Input.”
• Success/Error colour coding (green/red) for feedback.

Once you are happy with the interaction, add a web object to Adobe Captivate, select embed code, and paste the Code from Gemini into the embed code. Click outside the web object and then preview the project. If you need to make adjustments, use the Chat in Gemini to use regular English instructions to alter the code and thereby the output. Copy and paste the code into Captivate with each interaction.

You will need an API key to make this work. Visit aistudio.google.com/app/api-keys to get one.

Hi Maria,

Captivate 13.0’s AI features are designed for content creation and not for learner-facing interaction. Captivate’s AI is currently focused on generating slide content, Rewriting or improving text and assisting authors during design time. Adobe Captivate cannot generate passwords or logic based on learner input.

If you want this feature to be included in Adobe Captivate, please provide feedback to our product team. Open Adobe Captivate > Go to Help menu > Click on Provide Feedback > Enter the details > Submit the feedback.

Thanks,
Adobe Support