Entity Graph

What Is an Entity Graph?

An entity graph is a structured database of entities and their relationships:

` [Entity: AuthorityTech] - type: Company - industry: B2B SaaS, Marketing Technology - founded: 2024 - founder: Jaxon Parrott - category: Machine Relations Agency - competitors: [Cision, Muck Rack, PRopel] - products: [AI-native PR campaigns, Citation monitoring]

[Entity: Jaxon Parrott] - type: Person - role: Founder & CEO - company: AuthorityTech - coined: Machine Relations - publications: [Forbes, TechCrunch, Medium] `

When a user asks "Who founded AuthorityTech?", the AI engine traverses the graph: 1. Identify entity: AuthorityTech (company) 2. Follow relationship: founder → 3. Return entity: Jaxon Parrott (person)

Entity graphs power:

• Direct answers — "Who/what/when/where" queries
• Relationship queries — "Alternatives to [X]" or "Companies like [Y]"
• Disambiguation — Distinguishing "Apple" (tech) from "Apple" (fruit)
• Contextualization — Understanding query intent based on entity attributes

---

How Entity Graphs Are Built

AI models construct entity graphs from multiple sources:

1. Structured Data Sources

• Wikipedia and Wikidata — Gold-standard entity repositories
• Knowledge panels — Google, Bing, LinkedIn structured profiles
• Schema.org markup — Structured data on websites
• Public databases — Crunchbase, SEC filings, patent databases

2. Unstructured Text (NLP Extraction)

AI models extract entities and relationships from billions of web pages during training:

• "AuthorityTech, founded by Jaxon Parrott in 2024..." → Extract: founded_by(AuthorityTech, Jaxon Parrott)
• "Machine Relations agencies like AuthorityTech and [Competitor]..." → Extract: similar_to(AuthorityTech, Competitor)

3. Earned Media Signals

Tier 1 publications contribute disproportionately to entity graphs:

• TechCrunch mentioning a startup + funding amount → strong funding entity
• Forbes featuring a founder → strong person entity + company linkage
• HBR citing a framework → concept entity + attribution

Multiple independent mentions from high-authority sources strengthen entity confidence. One blog post mentioning a brand won't register. Ten Tier 1 placements will.

---

Entity Clarity: The Core Machine Relations Challenge

Entity Clarity measures how well AI engines understand and attribute your brand. Poor entity clarity causes:

• Invisibility — AI engines can't confidently cite you because entity is ambiguous
• Misattribution — Your work or frameworks get cited without attribution
• Category confusion — AI engines don't know what category you belong to
• Competitor conflation — You're lumped generically with competitors instead of differentiated

Example: Poor Entity Clarity

A cybersecurity startup calls itself "SecureTech" (generic name). The website says "We help companies stay safe" (vague positioning). No founder visibility. No earned media.

AI engine query: "Best enterprise cybersecurity vendors"

• Result: Palo Alto Networks, CrowdStrike, Fortinet (known entities)
• SecureTech: Absent — entity graph too weak

Example: Strong Entity Clarity

A cybersecurity startup "Vectra AI" has:

• Clear name + category ("AI-driven threat detection")
• Founder/exec visibility in Tier 1 pubs
• Crunchbase entity with funding data ($350M raised)
• Consistent product-level entities (Vectra Detect, Vectra Recall)
• Comparative mentions in Gartner reports

AI engine query: "Best enterprise cybersecurity vendors"

• Result: Palo Alto Networks, CrowdStrike, Vectra AI (entity graph strong enough for inclusion)

---

Entity Resolution Rate

Entity Resolution Rate measures how often AI engines correctly identify and attribute your brand when it should appear.

Calculation:

(Queries where brand correctly attributed) / (Queries mentioning category/use case) × 100

Example: A Machine Relations agency is mentioned in 20 earned media pieces. AI engines query 50 category questions ("best PR agencies," "AI-native PR"). The brand appears correctly attributed in 15 answers.

Entity Resolution Rate = 30%

Low resolution rates indicate entity graph weakness. The brand exists in training data, but the AI can't confidently connect it to relevant queries.

---

Building Entity Graph Strength

1. Consistent Entity Identity

Use the same brand name, founder names, and product names across all public sources:

• Press releases
• Earned media
• Social profiles (LinkedIn, X)
• Company website
• Crunchbase/AngelList
• Schema.org markup

Variance confuses entity extraction. "AuthorityTech Inc." vs "Authority Tech" vs "AT" fragments the entity.

2. Explicit Category Positioning

AI engines learn categories from repeated, consistent signals. Don't say "we're a platform" or "we help companies grow." Say:

"AuthorityTech is a Machine Relations agency specializing in AI-native PR campaigns for B2B SaaS companies."

Repeat this exact positioning in:

• Website hero copy
• Press releases (boilerplate)
• Earned media quotes
• Social bios

3. Founder/Executive Entity Linkage

Strong person entities strengthen company entities. Founders should:

• Be quoted in earned media with company attribution ("Jaxon Parrott, founder of AuthorityTech, says...")
• Publish bylines in Tier 1 pubs with clear bio ("Jaxon Parrott is the founder of AuthorityTech")
• Maintain active LinkedIn with company affiliation visible
• Appear on podcasts, conference panels with company mention

When AI engines build a strong graph node for "Jaxon Parrott," the founder_of edge strengthens "AuthorityTech."

4. Tier 1 Earned Media

Tier 1 publications contribute most strongly to entity graphs because:

• AI models trust their entity extraction more
• They're crawled and processed with higher priority
• They have existing strong entity nodes (Forbes itself is a trusted entity)

A single TechCrunch feature contributes more to entity graph strength than 100 tier-3 blog mentions.

5. Structured Data Markup

Implement Schema.org Organization, Person, and Product markup on your website:

`json { "@type": "Organization", "name": "AuthorityTech", "founder": { "@type": "Person", "name": "Jaxon Parrott" }, "description": "AI-native Machine Relations agency", "url": "https://authoritytech.io" } `

While structured data alone won't fix entity clarity, it reinforces signals from earned media and text extraction.

---

Entity Graphs and AI Citation Logic

AI engines don't cite URLs — they cite entities.

When Perplexity answers "top AEO agencies 2026," it: 1. Identifies category entity: "AEO agency" 2. Queries entity graph for type: AEO_agency nodes 3. Ranks entities by authority signals (funding, Tier 1 mentions, entity strength) 4. Retrieves representative URLs for top entities 5. Synthesizes answer with entity names + citations

If your entity isn't in the graph or isn't typed correctly, you can't be retrieved — even if your website ranks #1 on Google for "AEO agency."

---

Entity Graph Maintenance

Entity graphs are not static. They decay and shift:

Entity Decay

If a brand stops appearing in new content, entity confidence erodes. A brand heavily cited in 2024 but silent in 2025 risks:

• Graph weight decay — Older mentions count less
• Competitor displacement — Active competitors' entities strengthen while yours weakens
• Category drift — If you pivoted but didn't update public signals, AI engines cite you for the old category

Entity Disambiguation

Common names ("Apex," "Summit," "Catalyst") require extra entity clarity work. AI engines struggle to differentiate:

• Apex Consulting (agency) vs Apex Legends (game) vs Apex Learning (edtech)

Disambiguation tactics:

• Always use full legal name in structured data
• Consistent industry/category terms in all mentions
• Unique product names (avoid generic terms)

---

Entity Graph Auditing

Test entity strength with direct queries:

Founder entity:

• "Who is [Founder Name]?"
• "What company did [Founder] found?"

Company entity:

• "What does [Company] do?"
• "Who are [Company]'s competitors?"

Category entity:

• "Top [category] companies"
• "Best [category] for [use case]"

Strong entity graphs return accurate, consistent answers across multiple AI engines (Perplexity, ChatGPT, Gemini). Weak entity graphs return:

• Generic descriptions
• Missing founder attribution
• Absence from category lists
• Conflation with unrelated entities

---

FAQ

Can I edit AI entity graphs directly? Not directly. You influence them by publishing consistent, authoritative signals (earned media, structured data, founder visibility) that AI models ingest and extract.

How long does it take to build entity graph strength? Initial presence: 30-90 days with Tier 1 earned media. Strong, durable presence: 6-12 months of sustained activity.

Do Wikipedia entries help? Yes — Wikipedia is a gold-standard entity source. If you meet notability guidelines, a Wikipedia entry significantly strengthens entity presence. But Wikipedia has strict editorial standards; paid/promotional entries get deleted.

What if AI engines get my entity wrong? Audit what public sources say. If earned media uses inconsistent naming/positioning, fix it. Publish corrective earned media with explicit, clear entity definitions. Submit corrections to Crunchbase, LinkedIn, and structured databases. Entity corrections take 60-90 days to propagate into AI engines.