Inspiration
Consumers now ask AI assistants for product advice instead of searching online. Brands like Tesla don’t know if, how, or why AI recommends their products. We wanted a way to see how AI thinks and help brands improve their visibility in AI answers.
What it does
The agent collects information about a product and its competitors, tests how multiple AI models rank them, and analyzes the reasoning behind those rankings. Then it gives easy-to-understand recommendations for brands to improve their chances of being recommended by AI.
How we built it
Created a set of structured queries for pricing, market trends, and product innovation. Used Tavily to search online sources that AI models commonly rely on. Ran the queries on different LLMs to see how each model ranks the products. Built an analysis engine to extract reasoning and generate recommendations.
Challenges we ran into
AI models sometimes gave different answers for the same question. Hard to figure out which online sources influence AI reasoning the most. Balancing automated analysis with clear, actionable recommendations.
Accomplishments that we're proud of
Built a working prototype that can test multiple AI models for product ranking. Created a knowledge map showing what information AI uses to make recommendations. Generated practical actions Tesla could take to improve AI visibility.
What we learned
AI rankings are influenced by specific knowledge signals, not just popularity. Consistency and credibility of online information matter a lot. Structured queries help focus AI testing and analysis efficiently.
What's next for Generative Engine Optimization (GEO)
Expand to more product categories beyond EVs. Track AI recommendation trends over time. Build a platform for brands to monitor and improve their AI visibility continuously. Use GEO insights to inform marketing, product strategy, and content creation.
Log in or sign up for Devpost to join the conversation.