Dr. Ayse Ozturk

AI “Team of Rivals” Approach

WSJ’s Alexandra Samuel describes how she uses AI to create a virtual “team of rivals,” channeling diverse and conflicting perspectives to test her ideas. This thoughtful approach helps with making clearer decisions by surfacing different expert opinions, critiques, alternatives, and creative solutions.

Exercise:

AI Team of Rivals for McItalian Espresso

A glass of McItalian espresso topped with foam, featuring the McDonald's logo, placed on a marble coaster in a modern café setting.

Goal

Learn how to use multiple perspectives (via role-play or AI prompts) to evaluate a new product launch strategy and practice critical, creative thinking in marketing decision-making. This exercise will also help with getting the best set of recommendations from an AI bot.

Scenario

McDonald’s is considering launching McItalian Espresso, a premium, authentic-style espresso drink. Your task is to evaluate whether and how McDonald’s should launch this product using a ‘team of rivals’ approach.

Step 1 – Open Your Chatbot

Choose an AI chatbot you have access to (ChatGPT, Claude, Gemini, etc.). Open a new conversation.

Step 2 – Create Your Expert Panel

  • You’ll be asking the chatbot to role-play a group of fictional experts, each with a distinct perspective on McDonald’s launching the McItalian Espresso.
  • Options: In selecting your panel of experts, you can use the set or subset below, you can create and use other experts with different personas, or you can guide the AI to generate a different set of relevant panelists.
    • Examples of roles/experts:
      1. Brand Loyalist – argues for consistency with McDonald’s existing brand identity.
      2. Italian Coffee Purist – skeptical about authenticity, defends espresso culture.
      3. Finance Analyst – focused on costs, pricing, and ROI.
      4. Health Advocate – concerned with nutrition, sugar, and caffeine.
      5. Gen Z Trend Spotter – emphasizes cool factor, social media buzz.
      6. Operations Manager – thinks about speed, training, and logistics.
      7. Competitor Analyst – looks at Starbucks, Dunkin’, and indie coffee shops.
      8. Global Marketer – evaluates how this product plays in different markets.

Sample starter prompt (or adapt it in your own words):
“You are a panel of experts with conflicting viewpoints. Each expert has a unique personality and specialty. Take turns debating McDonald’s plan to launch a new product called McItalian Espresso. Speak in character, be bold, disagree with each other, and explain why others are wrong.”

Step 3 – Synthesize Options

When the rivals have finished debating, develop a summary/synthesis of perspectives. You can ask the chatbot:

  • “Create a table of the main points in the arguments and counterarguments”
  • “Now summarize the discussion into 3–4 clear strategic options for McDonald’s.”

Step 4 – Reflect and Decide

Write a short reflection:

  • Which option do you favor and why?
  • Which rival perspective most changed your mind?

Step 5 – Debrief

Discuss:

  • What did the disagreements reveal?
  • How did hearing multiple perspectives affect your thinking?
  • How can marketers avoid “yes-man” thinking when making big product decisions?

Bonus: Further Questions for Critical Thinking

In setting up your “team of rivals” by creating the panelists using an AI bot, consider these:

  • Who are the virtual experts/customers the AI includes in a “relevant set” of panelists that you would not have thought to include?
  • Who does the AI miss that you wish it had included?
  • What differences do you see between the panels the AI generates when maximizing differences versus maximizing relevant expertise?
  • Which assumptions or biases might shape the AI’s criteria for selecting panelists?
  • How might the AI’s panel composition change if the context (e.g., industry, culture, or audience) were different?
  • How could the AI’s panel choices influence the perspectives and outcomes of the case?

Acknowledgements: This assignment is inspired by WSJ’s Alexandra Samuel, many thanks! The purpose is to be able to push for better outcomes by a wholesome perspective obtained through a synthesis of a variety of contradicting opinions. There are similar tools that can be used for this purpose. For example, the contradictory chatbot on PlayLab.ai is a playful chatbot that gives three contradicting perspectives to a given question in the prompt.