Geo-David

Inspiration

In the age of AI-generated search results, brands are losing control of their narrative. When ChatGPT, Perplexity, or Google's AI Overviews answer questions about your company, they often hallucinate facts, miss key differentiators, or fail to mention you entirely.

We built Geo David because David (the protein bar company) needed a way to understand and fix how AI systems perceive their brand—turning the “Goliath” of AI misinformation into actionable insights.

What it does

Geo David is a unified GEO (Generative Engine Optimization) platform with three powerful tools:

GEO Auditor

Searches the web for your brand's keywords
Analyzes whether you're mentioned in top-ranking content
Identifies missed opportunities (articles that should mention you but don’t)
Extracts author contact info
Generates personalized outreach emails

GEO Comparative Agent

Compares your website against competitors using AI
Identifies content gaps
Analyzes:
- Schema markup
- Data density
- Q&A formatting
Generates ready-to-paste JSON-LD structured data to improve AI visibility

GEO Fact-Checker

Simulates how AI systems answer questions about your brand
Compares AI answers to your official documentation
Detects hallucinations
Provides:
- Detailed remediation steps
- Prevention tips
- Suggested FAQ content to add to your website

How we built it

Frontend: React + TypeScript with Vite for lightning-fast development
AI: Google Gemini 3 (Flash & Pro models) for content analysis, question generation, and verification
Web Search: Tavily API for real-time search engine result analysis
Web Scraping: Firecrawl API for extracting clean markdown from any URL
Deep Research: Gemini's built-in search grounding for retrieving official ground truth
Markdown Rendering: react-markdown for beautiful content display

Challenges we ran into

Rate Limits: Balancing multiple concurrent AI calls while respecting Gemini's free tier limits required careful model selection (Flash vs Pro) and sequential processing
Hallucination Detection: Defining what constitutes a hallucination vs. a minor inaccuracy required nuanced prompting
Author Extraction: Getting AI to reliably extract author names and emails from unstructured web content
JSON-LD Generation: Ensuring the AI generates syntactically correct, copy-paste-ready structured data

Accomplishments that we're proud of

One-click outreach: The Auditor generates personalized emails for missed opportunities with a single click
Ready-to-paste code: Comparative analysis outputs complete HTML that non-technical users can add directly to their site
Actionable remediation: The Fact-Checker provides step-by-step fixes and even generates exact FAQ Q&A to add
Unified experience: Three powerful tools in one tabbed interface with centralized configuration

What we learned

GEO is the new SEO: Traditional SEO focused on ranking—GEO focuses on being accurately represented in AI responses
Structured data matters more than ever: JSON-LD schema is becoming the language AI understands best
The best fix is prevention: Adding clear, factual content to your website is cheaper than fixing hallucinations after they spread

What's next for Geo David

Automated monitoring: Schedule recurring audits to track brand visibility over time
Competitor tracking dashboard: Visualize GEO metrics vs. competitors week-over-week
CMS integrations: One-click deployment of schema fixes to WordPress, Shopify, and Webflow
Multi-brand support: Agency mode for managing GEO across multiple client brands
Chrome extension: Highlight AI mentions (accurate vs. hallucinated) as you browse the web