Refora
AI visibility intelligence for cybersecurity brands
The Challenge
As AI search becomes the primary way buyers research products, cybersecurity companies have no visibility into how they're represented in AI responses. Traditional SEO tools don't track AI-generated content, leaving brands blind to a critical new discovery channel.
The Solution
I built a SaaS platform that queries multiple AI engines (ChatGPT, Claude, Gemini, Perplexity) with buyer-intent questions and analyzes the responses. The system calculates 'Share of Answer' metrics—how much of each response mentions your brand, where you appear, sentiment analysis, and citation tracking.
Key Features
Multi-LLM Integration
Unified AI client abstraction querying ChatGPT, Claude, Gemini, and Perplexity with graceful degradation if any engine fails.
Share of Answer Metrics
Calculates percentage of AI response dedicated to your brand, mention position, citation rates, and sentiment analysis.
Competitive Intelligence
Compare visibility metrics vs competitors, identify content gaps where competitors appear but you don't.
Automated Alerts
Visibility drop/surge detection, competitor surge alerts, negative sentiment spikes with Slack/email notifications.
The Outcome
The platform provides actionable intelligence on AI visibility, enabling brands to identify content gaps where competitors appear but they don't, track visibility trends over time, and receive automated alerts for significant changes. Multi-tenant architecture supports per-organization data isolation.
Engineering Highlights
- Designed event-driven architecture with Inngest for fault-tolerant background job processing
- Built AI engine abstraction layer with unified interface and engine-specific adapters
- Implemented multi-tenant architecture with per-organization data isolation
- End-to-end type safety from API routes to database queries with Drizzle ORM