How this Yelp AI PM works backward from “golden conversations” to create high-quality prototypes

Episode Overview

• Product Manager Priya Matthew Badger shares her novel workflow for managing AI products, which involves handling both the user interface and the non-deterministic back-end logic.
• The process starts with writing "golden conversations"—ideal user-AI interactions—which serve as a text-based prototype and the foundation for product requirements.
• This conversational-first approach is demonstrated by using example conversations and UI mockups as a prompt to generate a functional prototype with AI tools like Claude.
• The workflow concludes by using AI for rapid design exploration, generating multiple differentiated UI/UX flows to accelerate ideation and compress the development cycle.

Key Concepts

• Golden Conversations: The practice of starting AI product design by writing out an ideal, example conversation between a user and the AI. This defines the target experience and serves as the product's foundation, allowing teams to work backward from a clear goal.
• PRD as a Prompt: Using a combination of example conversations, system instructions, and visual mockups as a comprehensive Product Requirements Document (PRD) to prompt an AI to generate a functional application.
• Rapid Prototyping and Iteration: Leveraging AI tools to quickly build and test interactive prototypes, moving from a text-based idea to a tangible, shareable mockup to get a realistic feel for the user experience.
• AI for Design Exploration: Using AI not just for execution but as a creative partner to generate multiple, distinct UI/UX solutions for a single feature, thereby accelerating the ideation process and exploring a wider range of possibilities.

Quotes

• At 0:31 - "We start with golden conversations, what's the experience that you're trying to drive?" - Priya introduces her method of using ideal user-AI conversations as the starting point for product design.
• At 17:14 - "So you really are using example conversations and just reference designs as your PRD here..." - The host summarizes Priya's novel approach of using conversational examples and UI screenshots as the input for generating a functional app.
• At 19:16 - "Sometimes a response that looks fine when you have it in a doc, feels really long when you see it in like the little chat bubble in the mobile interface." - Priya highlights the value of interactive prototypes, as they provide a more realistic feel for the user experience compared to static documents.
• At 26:15 - "Give me some options for how the 'Start with a photo' flow could work to make it feel more guided for the user. Think expansively and make each option differentiated." - Priya prompts an AI to generate multiple distinct design ideas, demonstrating how to use AI for creative exploration.
• At 31:28 - "You're starting from a kind of example consumer experience first. You're backing into a rough prototype... And then you're using that as a jumping-off point to fork and inspire a couple different versions..." - The host recaps the entire AI-native workflow, emphasizing its non-linear and iterative nature.

Takeaways

• Begin AI product design by scripting "golden conversations" to anchor the entire development process in the ideal user experience.
• Treat a collection of example conversations, system instructions, and reference designs as a modern, dynamic PRD to generate functional prototypes.
• Move beyond static documents and embrace interactive prototypes to accurately test and feel the flow and tone of AI-driven conversational interfaces.
• Leverage AI tools for creative exploration by prompting them to generate multiple, differentiated design options, which can significantly accelerate ideation.