Cross-referenced analysis across Wikipedia, Wikidata, Google’s Knowledge Graph, and industry-specific sources. These are the grounding datasets LLMs treat as authoritative — the entities present here shape the entities LLMs surface in responses.
LLMs don’t invent facts about brands. They pull from a small number of trusted structured sources. GEO Jetpack analyses your presence in each — and scores that presence using our KGIS model.
Article presence, article quality (stub through featured), and section coverage. LLMs treat Wikipedia articles as high-authority grounding when they exist.
Entity presence, claim completeness, and property coverage. Wikidata is the highest-weighted source because it’s structured and LLMs consume it directly.
Google Knowledge Graph presence, entity type, and description quality. Feeds Google’s AI Mode and Gemini responses directly.
Sector-specific knowledge sources — Crunchbase for tech, MusicBrainz for music, Companies House for UK entities. Weighted by industry relevance.
The KGIS (Knowledge Graph Inclusion Score) aggregates presence across all four sources into a single 0-100 score, weighted by LLM grounding importance. Track it over time as you close the gaps.
The gap analysis compares your entity coverage against named competitors across all four sources — highlighting the specific claims, articles, and properties they have that you don’t.
When an LLM is asked “what does Aviva Investors do,” it doesn’t generate an answer from nothing. It grounds the response in structured facts pulled from Wikipedia, Wikidata, Google’s Knowledge Graph, and — for financial services specifically — regulatory and Companies House data.
If your brand’s entity is thin or absent in those sources, the LLM has less to ground on. It may skip you, describe you in less detail than competitors, or — worst case — hallucinate details based on adjacent entities. Improving your knowledge graph presence isn’t optional for AI Search visibility. It’s the foundation.
The paradox is that many marketing teams optimise heavily for their own site content while ignoring the third-party sources LLMs actually treat as authoritative. That imbalance is why so much SEO effort produces flat AI Search results — you’re pulling the wrong lever.
You cannot content-optimise your way to strong AI Search visibility if your knowledge graph presence is weak. The two run in parallel.