How custom GPTs can make you a better manager | Hilary Gridley (Head of Core Product at Whoop)

How I AI How I AI May 18, 2025

Audio Brief

Show transcript
This episode explores how managers can scale their leadership and expertise using custom GPTs to provide consistent feedback and develop personalized AI coaches for their teams. There are four key takeaways from this conversation. First, codify your expertise by training AI with before and after work examples. This method reverse-engineers your judgment, teaching the AI your specific standards and thought process. It enables AI to deliver consistent, high-quality feedback, automating initial reviews and freeing managers for strategic tasks. Second, leverage meta prompting to efficiently build robust custom tools. Engage an AI in a defining conversation, then instruct it to synthesize that entire discussion into a comprehensive final prompt. This technique streamlines custom GPT creation, saving time and ensuring precise tool design. Third, create personalized AI coaches tailored to individual team members. Moving beyond generic advice, these tools address specific developmental needs and career goals. An AI coach can help an employee practice anticipating stakeholder questions or refine their data visualization skills, fostering hyper-specific growth. Fourth, utilize AI as a private sparring partner to sharpen your own thinking. Ask it to pressure-test ideas, identify blind spots, and strengthen strategic arguments. This allows for clear articulation of a point of view before wider presentation, significantly enhancing your contribution to strategic discussions. By applying these advanced AI techniques, managers can significantly enhance their effectiveness, elevate team performance, and foster a more engaged and empowered employee experience.

Episode Overview

  • Learn how to scale management by creating custom GPTs that act as a "digital clone" of your expertise, providing consistent and high-quality feedback to your team.
  • Discover the process of "reverse-engineering" your own judgment by feeding an AI "before-and-after" examples of work to teach it your specific evaluation criteria.
  • Explore the power of "meta-prompting," a technique where you ask an AI to synthesize your conversation into a robust, final prompt for a new custom tool.
  • Understand how to move beyond generic tools to create personalized AI coaches tailored to the unique growth areas of individual team members.

Key Concepts

  • AI as a Managerial Force Multiplier: Managers can build custom GPTs to automate initial feedback and coaching on recurring tasks (like slide deck reviews), freeing them up for higher-level strategic work while empowering their team.
  • Reverse-Engineering Your Judgment: The core method involves training an AI by providing it with "good" and "bad" examples of a work artifact. The AI then deduces the underlying principles and criteria, which the manager can iteratively refine.
  • Meta-Prompting: Instead of writing a complex final prompt from scratch, a manager can have a conversation with an AI to define the criteria and then ask the AI to synthesize that entire conversation into a structured, comprehensive prompt for a new custom GPT.
  • Personalized Employee Development: The true power of this approach is creating hyper-specific AI tools for individuals. A generic "Slide Clarity Coach" can be adapted into a personalized coach that helps a specific employee practice anticipating stakeholder questions or improving their data visualization skills.
  • AI as a Sparring Partner: Use AI not just to get feedback, but to challenge your own thinking. Ask it to identify blind spots, test the strength of arguments, and help you articulate a point of view more clearly before sharing it with a wider audience.

Quotes

  • At 4:15 - "I think of this as kind of like reverse engineering a recommendation algorithm for yourself and for your own preferences, which is just starting to collect examples of good and bad." - Hillary describes her core technique for teaching an AI her specific judgment by providing it with before-and-after examples.
  • At 6:30 - "YOUR job is to write the prompt for it." - Hillary demonstrates a powerful meta-prompting technique where she asks the AI to create the final, structured prompt for a new custom GPT based on their conversation.
  • At 14:58 - "What an improved employee experience. You know, you have this frazzled boss who's busy that gives you quick feedback... versus this like endlessly patient and clearly articulate, always available version of your manager." - Karen highlights the benefit of AI-powered coaching from the employee's perspective.
  • At 24:06 - "The way you get pulled is by showing evidence of being able to contribute to strategic thought and move it forward in a way that's really helpful. And the easiest way to do that is... write up your point of view." - Hillary's advice on how to get invited to more important meetings, a task AI can help with.
  • At 34:06 - "I go like Mean Girls on them. Like, I sell them out to their friends. I go over to Claude and I'm like, 'Claude, you will not believe...'" - Hillary shares her humorous and surprisingly effective technique for getting AI models to follow instructions by complaining to a competing model about the other's failures.

Takeaways

  • Codify your expertise by providing an AI with "before-and-after" examples of your work to teach it your specific standards and thought process.
  • Use "meta-prompting" by asking the AI to synthesize your feedback and criteria into a final, robust prompt for a custom GPT, saving time and creating a better tool.
  • Create personalized AI coaches for your team members that are tailored to their specific developmental needs and career goals.
  • Leverage AI as a private "sparring partner" to pressure-test your ideas, identify blind spots, and strengthen your strategic arguments before presenting them.