The secret to better AI prototypes: Why Tinder’s CPO starts with JSON, not design | Ravi Mehta (product advisor, previously EIR at Reforge)

Data-driven prototyping bridges the gap between high-level conceptual ideas and functional product design by prioritizing structured data schemas over vague, vibe-based prompts. By defining data models—such as JSON structures for trip itineraries—before generating UI, product teams achieve greater consistency and accuracy in their prototypes. Integrating external tools via the Model Context Protocol, such as Unsplash for authentic media, eliminates common AI hallucinations like broken links or contextually inappropriate images. Furthermore, elevating image generation in tools like Midjourney requires a shift toward photographic terminology, specifically defining subject, setting, and style through camera and film metadata. Ravi Mehta, CPO at Tinder, emphasizes that while AI significantly lowers the barrier to technical craft, success remains rooted in human taste and a deep understanding of consumer needs, ensuring that prototypes serve as effective tools for testing and iteration.