How we map 5,600+ companies across 30 verticals using entity classification, relationship edges, structured capabilities, and AI visibility scoring.
Unlike flat company directories, geo.sig.ai organizes the technology landscape as a knowledge graph — a network of typed entities connected by explicit relationships. Every company, product, and platform in our database is classified by what it is, what it does, who it competes with, and how it connects to the broader ecosystem.
This structure powers our AI visibility scoring, competitive analysis, and the brand intelligence pages that help companies understand their position in AI-powered search results.
Every entity in the graph is classified into one of six types. This determines how it appears in rankings, how its AI visibility score is computed, and what Schema.org structured data is emitted for search engines.
Legal entities and parent organizations
e.g., OpenAI, Microsoft, Salesforce
Market-facing identities (default classification)
e.g., Stripe, Datadog, HubSpot
Specific software offerings of a parent company
e.g., ChatGPT, Claude, Gemini
Developer ecosystems and cloud infrastructure
e.g., AWS, Azure, Google Cloud
AI foundation models
e.g., GPT-4, LLaMA, Mistral
Open-source projects and frameworks
e.g., LangChain, Next.js, PyTorch
Entities are connected by typed, directional edges that capture real-world relationships — competitive dynamics, ownership, technology dependencies, and integrations. Symmetric relationships (like competition) are stored in both directions.
Direct market competitors in the same category
ChatGPT competes with Claude
A product or service offered by a parent company
ChatGPT is a product of OpenAI
An organization owned by a parent company
AWS is a subsidiary of Amazon
Products that work together via API or partnership
Slack integrates with GitHub
Uses another entity's core technology
GitHub Copilot is powered by OpenAI
Strategic financial investment
Microsoft invested in OpenAI
Corporate acquisition
Slack was acquired by Salesforce
Core dependency or foundation technology
Vercel is built on Next.js
Each entity is tagged with structured capabilities — what it actually does — organized by domain (vertical). This replaces free-text categorization with a controlled vocabulary of nearly 3,000 capabilities, enabling precise searches like "find all code generation tools in Developer Tools."
Capabilities are assigned a proficiency level: primary (core function), secondary (additional capability), or emerging (new/experimental feature).
Every entity receives an AI Visibility Score (0–100) estimating how often major AI systems — ChatGPT, Claude, Gemini, Perplexity, and Grok — mention the brand when users ask relevant questions in its category.
Scores are computed from market position signals including industry tier, category rank, brand authority, and cross-platform consensus. Scores are labelled "Estimated (Beta)" while we build real-time AI query tracking. Trend direction (rising, declining, stable) tracks score trajectory over a 12-week window.
Brand profiles are enriched from 20+ external data sources, cross-referenced for accuracy. No single source is trusted in isolation.
The knowledge graph is accessible via multiple interfaces:
Brand Graph API/api/public/brands/brand/{slug}/graphEntity type, parent/children, edges, and capabilities
Brand Profile API/api/public/brands/brand/{slug}Full brand detail with enrichment data
MCP Serverhttps://geo.sig.ai/mcp7 tools for Claude, Cursor, Windsurf agents
Markdown Profiles/brands/{slug}.mdLLM-optimized markdown for each brand
OpenAPI Spec/api/public/openapi.jsonFull API specification for auto-discovery
Browse brand profiles to see entity types, competitive edges, and AI visibility scores in action.