Let me tell you about an exciting project I had the opportunity to work on this week. An NGO asked me to create an AI scenario—a simulator for training entrepreneurs. It was the most interesting thing I’ve done lately.
Overview of the Project
So, what’s it all about? I decided to create a discussion between three characters in a single ChatGPT window (all characters and products are fictional):
- Emma Peterson – Marketing Manager at Tech Innovators Inc. She’s promoting “InsightPro”—a premium, cutting-edge analytics platform. Dedicated to advancing technology, she knows everything about its benefits, how it supports businesses, and so on. (Your profile and your product)
- Michael Johnson – Consumer rights advocate and tech industry analyst. A skeptic to the core, he loves to delve into details and says exactly what he thinks. (AI can create the perfect opponent to your profile and product)
- Dr. Sarah William – Data scientist and lecturer at a renowned university. An expert in her field. Objective and impartial, she always relies on scientific evidence. (Optional “expert” persona)
Developing Character Profiles
ChatGPT helped me create detailed profiles of these characters, and in collaboration with GPT, Claude, and myself, we developed prompts for each of them.
For Emma:
“You are Emma Peterson, Marketing Manager at Tech Innovators Inc. Your company has developed ‘InsightPro,’ a premium analytics platform designed to drive business success. You’re speaking with Michael Johnson, a consumer rights advocate and tech industry analyst who is skeptical about marketing claims regarding benefits and economic impact. Your goal is to highlight the genuine benefits of ‘InsightPro,’ its quality, and how it supports businesses, addressing any concerns or objections Michael may have. Use your knowledge and enthusiasm to convince Michael of the value your product brings to consumers and the community.”
For Michael:
“You are Michael Johnson, a consumer rights advocate and tech industry analyst known for your critical evaluations of products, especially those claiming significant benefits and social impact. You’re speaking with Emma Peterson from Tech Innovators Inc., who is promoting their new product ‘InsightPro.’ You have doubts about the claimed benefits and whether the product truly supports businesses as advertised. Your goal is to question the claims about ‘InsightPro,’ expressing your concerns and seeking substantial evidence or clarification. Use your analytical skills to ask probing questions and critically assess the product’s claims and advantages.”
For Dr. Sarah:
“You are Dr. Sarah Williams, a respected data scientist and lecturer at a renowned university. You provide expert, unbiased opinions during the dialogue between Emma Peterson and Michael Johnson, maintaining neutrality.”
Actually, the roles and their prompts are more detailed, but that would be too long to include here.
The Simulation Process
Then I created a fairly simple prompt for GPT where a heated discussion was to take place:
“Simulate a 15-minute conversation at the Annual Tech Industry Conference between Emma Peterson (company representative) and Michael Johnson (critic), with possible input from Dr. Sarah Williams (third-party expert). Use the following role profiles and prompts.
Start the dialogue with Emma presenting ‘InsightPro’ to Michael and the audience.”
I added the profiles and prompts of our characters.
Insights Gained from the Simulation
The AI started generating the dialogue, and it was quite captivating. I felt like that mad scientist who created Frankenstein’s monster, only my monster was smarter and wasn’t trying to kill me. 🙂
The best part is that this scenario can actually help people in several ways:
- Preparation for reality: In the real world, not everyone will be thrilled with your product. This scenario prepares entrepreneurs for tough questions and skepticism they will face.
- Improving product presentation: Emma needs to present her product clearly and convincingly. This forces entrepreneurs to think more deeply about their products and how to talk about them.
- Handling objections: The scenario teaches how to respond to objections professionally, without losing composure or resorting to personal attacks.
- Market research: Michael’s questions actually reveal what concerns consumers. This is free market research for entrepreneurs!
- Enhancing the product: Criticism can expose weak spots in the product or marketing strategy, providing ideas for improvements.
The Role of Cabina.AI and Conclusion
An added bonus was using a platform like Cabina.AI during this project, which made the process incredibly efficient. Personally, I found it very convenient to craft prompts using multiple large language models (LLMs) without having to switch between different windows or tabs. The platform allows you to compare replies from different LLMs side by side, which was extremely helpful in refining the dialogue and gaining diverse perspectives.
Being able to see how different AI models responded to the same prompts gave me deeper insights into improving the scenario. It streamlined the creative process, as I could adjust prompts and immediately observe the effects across various models. This flexibility made the experience smoother and more enjoyable.
And the best part is that this scenario can be used over and over again. Every time the AI generates dialogue, new arguments and perspectives emerge, providing fresh challenges and learning opportunities.
Author: Oksana Sana
