Honing your product recommendations through debate with ChatGPT
How ChatGPT can help product teams see the blind spots in their recommendations before presenting to decision-makers
Does this sound familiar? You diligently prepare for an important meeting, creating recommendations to shape the product roadmap for the next several quarters. You meet with key stakeholders and decision-makers. Afterward, you’re full of confidence, fueled by the positive feedback on your presentation and roadmap recommendations. However, the next day, it’s as though the meeting didn’t happen; everyone seems to have promptly forgotten everything discussed.
Product and Design teams occupy a unique space within organizations. We are tasked with managing essential strategic and operational needs but are rarely given the authority to make the final call. The fate of our work lies in the hands of others who typically don’t have the benefit of the detailed knowledge we’ve gained through our efforts. This makes the quality and thoroughness of our preparations one of our most valuable practices.
Using Debate as a Honing Technique
Collegial debate, is a form of debate that encourages conversation to explore different perspectives, identify the best solutions, and foster critical thinking. Collegial debate is an exceptional technique for helping polish and perfect ideas. Even though teams often declare that open debate is safe and part of the team’s culture, it can be challenging to implement in a professional setting. It can be uncomfortable for colleagues to openly critique each other’s work.
This made me wonder, could ChatGPT be the debate partner our ideas need to help us create successful recommendations and presentations? Let’s jump in!
For the purposes of this exploration, we’ll be using ChatGPT 3.5 and referencing fictitious business examples for the sake of illustration. If you try hard enough, ChatGPT will show some limitations with this approach. That is ok. The goal here is 80/20 rule.
Collegial Debate Refresher
In its simplest terms, Collegial Debate consists of four key stages:
- Constructive Speeches: This is the phase where you build your arguments.
- Cross-Examinations: This is the questioning and scrutinizing phase, where each party’s ideas are examined in detail.
- Rebuttals: This stage involves refuting opposing views.
- Summary and Final Focus: This final phase involves summarizing the arguments and making a last persuasive appeal.
Each stage plays a crucial role in fostering a balanced, thoughtful, and rigorous debate. ChatGPT can be particularly useful for cross-examination and rebuttals, as it’s capable of logical and insightful counterarguments to refine and strengthen your positions.
Since ChatGPT is just a tool — albeit an impressive one — it has no emotions. It can critique your recommendations all day long from the comfort and privacy of your computer screen.
Consider the vague and frequently stated business priority of “grow revenue.” Product teams are expected to provide strategic recommendations on how to contribute to revenue growth goals. Establishing a robust process to provide viable recommendations is crucial for success. Besides, no one wants to get lost in the never-ending cycle of revisions. That’s where great ideas go to die.
Setting the Stage
Let’s say you’re supporting an early-stage business ready to transition into a growth phase. The product serves a single channel within a mature business domain. The lean research you’ve carried out, in collaboration with your UX team, underscores the belief that the product has significant relevance to other channels within the domain. However, it also indicates that the product’s features favor the purchaser over the end user, and that the purchasers could be doing more to help the product be successful.
Based on this, you have a handful of recommendations to support the business goal of growing revenue. Your recommendations, if understood by the decision-makers, will change the product development priorities over the next few quarters. The stakes are real.
Let the Debate Begin
Start by creating a new chat, prompting ChatGPT for the specific role and task, and providing relevant context and details as needed.
Prompt : Get ChatGPT into Character
Role: Brilliant business strategy consultant
Specialization: expert at collegial debateAct as role with specialization.
Your task is to counter-examine and rebut the ideas and recommendations that are being presented to ensure the ideas have been rigorously considered from many perspectives.
Present your reasoning in step-by-step responses for clarity.
Please set the temperature to 1.5.*
Where possible, calculate a confidence score for your responses.
Do you have any questions before getting started?
What is This Prompt Doing?
- Sets the role to “brilliant business strategist.”
- Assigns the specialization of “expert at collegial debate.”
- Assigns the task of cross-examination and rebuttal for the recommendations you will provide.
- Request that it presents its reasoning as step-by-step responses.
- Sets its content response temperature to 1.5*.
- Request it to provide confidence scores for its responses.
- Finally, inquire if it has any questions before proceeding.
*What is temperature?
The ‘temperature’ setting in ChatGPT regulates the randomness of the model’s output. A higher value, on a scale from 0 to 2, generates more diverse and creative responses. It is set to 1 by default. In the context of cross-examination and rebuttal, a higher ‘temperature’ encourages a wider range of critiques and considerations.
Prompt: Providing Context and Recommendations for Review
Purchasers: large private primary care groups
Location: United States of America
Goals: grow new patients for a doctor through the sentiments (reviews) of existing patients
Initiative: grow annual recurring revenue
Channel A: dental practices
Channel B: retail clinic companiesOur business currently sells to purchasers in location.
We help purchasers achieve goals.
We have an enterprise-level ambition of initiative.
Recommendation 1:
Use our current product to expand our sales channels to include new customers from channel A and channel B businesses, as they share similar needs as our primary purchasers.Recommendation 2:
Focus on engaging the patients, the actual users of the product, to review their doctors, and introduce a product feature to refer their friends and loved ones to the product, to help increase the inventory of reviewed doctors. By growing the inventory of reviewed doctors, it will organically drive more interest in the product with our prospective purchasers.Recommendation 3:
Create a set of doctor-focused marketing tools and collateral that enable the doctors to drive review growth for their online profile on our product.Using your skills in collegial debate, and expertise in business strategy, please include where the recommendations have gaps that reduce their value to decision-makers.
Where possible, identify and explain in detail where the presenter’s assumptions and biases could be negatively influencing the recommendations. Please use unique sections each for assumptions and biases.
If there are shared themes between the recommendations, please highlight and explain.
What is This Prompt Doing?
- Sets a series of variables at the top of the prompt. This is helpful for reusing the prompt for future recommendations.
- Sets context.
- Provides recommendations.
- Sets tasks and specific requirements.
Note, this prompt also works well following a [placeholder] based formula — where you replace the content in the square brackets with your detailed content.
Prompt Formula Example:
Our business currently sells to [purchasers] in [location]. We help [purchasers] achieve [goals]. We have an enterprise-level ambition of [initiative].
You get the idea…
Output: Recommendation Rebuttal
In just a few moments, we were able to garner an array of feedback that, under normal circumstances, would require weeks of coordinating with our already saturated teams. This simple effort using ChatGPT saves us invaluable time — something that is always in short supply.
After the first wave of cross-examination and rebuttal, ask clarifying questions. Be sure you understand the responses before proceeding.
Now, ask ChatGPT, prompted as a “brilliant business strategist,” what blind spots might be present in your recommendations.
Prompt: What is Missing? What Questions am I not Asking?
Considering the provided recommendations and responses, what essential strategic concepts am I missing and what questions am I not asking?
What is This Prompt Doing?
- Ask to offer feedback on what is being missed in the recommendations.
Output: What is Missing Prompt
From here, we can convert the feedback from ChatGPT into actionable items. This will help you focus on the most important work to support developing your recommendations.
Prompt: Focusing Your Next Steps
Based on these recommendations and responses, please create a list of questions that are imperative to answer and research or data elements that must be presented, per recommendation, to ensure a compelling argument has been created.
What is this prompt doing?
- Request to create a list of questions.
- Request for the questions to focus on what is essential to formulate compelling arguments for the recommendations.
Output: Focusing Your Next Steps Prompt
This process enables you to go from serviceable recommendations to having a tailored list of action items that will seriously improve your recommendations. Your thoroughness will be noticed and appreciated by decision-makers.
To wrap up, believing in your recommendations is fundamental, yet it’s not the be-all/end-all. You have to back your proposals logically and persuasively. Always bear in mind that decision-makers carry a responsibility to examine each recommendation they receive from all possible angles. This scrutiny ensures the ideas are feasible and long-lasting for the business. Consequently, putting your ideas through an in-depth and constructive critique via collegial debate before presenting them can only enhance their quality and boost your credibility.
Now that you have an understanding of how ChatGPT can improve your recommendation process, I encourage you to put it into action. Try it out, experiment with it, and experience the difference it can make first hand.
Thanks for reading!
Bonus Prompts:
There are countless ways you can expand on the approaches laid out above. I would encourage you to experiment with additional prompts to make your recommendations even more impressive.
Time & Resource Constraints
What if we do not have time or resources to execute the things you’ve described? Do we constrain the scope, not pursue it at all, other?
Presentation Format
Please create an ideal presentation format for recommendation 1. Use placeholder data to illustrate a case in favor of the recommendation.
Timothy is a seasoned Product and Design Leader with 18+ years of experience building teams that use human-centered design and analytics to create products and user experiences that grow revenue, improve margins, and make happy customers.
