The 40 Citation Surfaces That Determine Your LLM Visibility

Generative engine optimization is fundamentally an off-site discipline. The question is not "what should be on my website?" — it is "where else does my brand exist in a form an LLM can cite?" This guide maps the 40 surfaces that answer that question.

Why 40 surfaces — not 5, not 100

VISIBLD's audit methodology converged on 40 after analyzing citation patterns across 12,000+ LLM responses. Below 40, the audit misses category-specific surfaces (e.g., GitHub for developer tools, AppExchange for Salesforce ISVs). Above 40, the surfaces have negligible citation weight and audit time exceeds value delivered. The 40 surfaces represent ~92% of total citation volume across the 8 major LLMs.

Source: VISIBLD citation analysis, 12,000+ LLM responses across ChatGPT, Perplexity, Claude, Gemini, Copilot, Grok, DeepSeek, and Google AI Overviews. May 2026.

Tier 1 — Entity anchors (5 surfaces)

Entity anchors are the canonical sources LLMs use to disambiguate your brand. Without consistent entity anchors, an LLM may fail to identify your company at all, or conflate it with a similarly-named entity. These surfaces have the highest citation weight per unit of effort.

• Your website — primary identity surface. Must have Organization schema, founder Person schema, and consistent NAP (name, address, phone).
• Wikipedia — 47.9% of ChatGPT citations. Highest single-surface weight in B2B GEO.
• Wikidata — provides the Q-ID that connects your entity across Wikipedia, Google Knowledge Graph, and several LLM training pipelines.
• Crunchbase — funding, founders, employee count, category. Heavy weight in Copilot and Gemini.
• LinkedIn company page — employee verification, leadership profiles, content velocity. Disproportionate weight in Copilot.

Tier 2 — Software review cluster (6 surfaces)

The review aggregator cluster is the most heavily-weighted Tier in B2B SaaS GEO. Comparison queries ("best CRM for X," "Salesforce vs HubSpot") pull citations almost exclusively from this Tier.

• G2 — dominant weight in Gemini and Copilot for B2B comparisons.
• Capterra — strong weight for SMB and mid-market software queries.
• Gartner Peer Insights — enterprise validation; Copilot weights heavily.
• TrustRadius — Claude weights TrustRadius for analytical buyer queries.
• Software Advice — secondary aggregator; cited in long-tail vertical queries.
• AlternativeTo — captures "alternatives to X" query class across all 8 LLMs.

The audit checklist per surface: profile completeness, screenshot quality, feature coverage, response to negative reviews, recency of new reviews (30/60/90-day velocity), and category placement.

Tier 3 — Technical & developer (5 surfaces)

Disproportionate weight for any category with technical or developer-adjacent buyers. DeepSeek pulls almost exclusively from this Tier.

• GitHub — repository presence, stars, fork count, README quality, issue activity.
• Stack Overflow — questions and answers tagged with your product. Cannot be authored — must be earned.
• Hacker News — Show HN posts, comment threads, "Ask HN" mentions. High Perplexity weight.
• Hugging Face — for AI/ML categories. Model cards, datasets, organization profile.
• Spiceworks — IT community; cited in MSP and infrastructure queries.

Tier 4 — Press & wire (6 surfaces)

Tier-1 press provides the longest-half-life citation lift in GEO. A single TechCrunch or Bloomberg mention propagates through training data refreshes for 12–18 months.

• PR Newswire — distribution to news aggregators and LLM training crawlers.
• BusinessWire — equivalent reach; preferred for funding and earnings news.
• TechCrunch — highest single-domain weight in B2B SaaS press citations.
• Bloomberg — financial credibility; cited in enterprise and IPO-stage queries.
• Reuters — international weight; cited globally across all 8 LLMs.
• Wall Street Journal — premium signal for enterprise and growth-stage companies.

Tier 5 — Community signals (4 surfaces)

Community signals are the highest-leverage and least-controllable surfaces. Perplexity in particular weights these heavily.

• Reddit — 46.7% of Perplexity cited sources. Subreddit relevance and comment depth matter more than upvote count.
• Quora — declining relevance but still cited in long-tail buyer-question queries.
• Product Hunt — launch-day weight is high; trail-off after 30 days but profile remains a citation surface.
• Podcast show notes — emerging surface. LLMs increasingly index podcast transcripts via Apple, Spotify, and YouTube.

Tier 6 — Ecosystem directories (5 surfaces)

• Salesforce AppExchange — for ISVs and Salesforce ecosystem.
• AWS Marketplace / Google Cloud Marketplace — cloud-native software discovery.
• Zapier app directory — integration-led GEO; Zapier listings are heavily indexed.
• Microsoft AppSource — Office 365 and Azure ecosystem placement.
• App stores (iOS / Google Play) — for B2B SaaS with mobile components.

Tier 7 — Firmographic databases (5 surfaces)

Firmographic surfaces drive trust and verification signals. Often overlooked because they are not "marketing" channels — but LLMs query them extensively to verify company facts.

• PitchBook — funding, valuation, cap table accuracy.
• CB Insights — categorization, market sizing, named competitor lists.
• Glassdoor — employee verification, culture signals, leadership ratings.
• AngelList / Wellfound — early-stage credibility, team verification.
• Job boards (LinkedIn Jobs, Indeed) — hiring velocity is a growth signal LLMs surface.

Tier 8 — Content distribution (4 surfaces)

• YouTube — transcripts indexed; product demos and explainers cited in how-to queries.
• Medium / Substack — long-form thought leadership; cited heavily by Claude.
• Trustpilot — sentiment surface; cited in reputation and "is X legit" queries.
• Built In — culture and hiring; tertiary but cited in employer-brand queries.

Weighting per LLM

The 40 surfaces are not weighted equally across LLMs. Approximate citation share per platform:

Source: VISIBLD platform-specific citation analysis, May 2026. Distributions vary by category; B2B SaaS averages shown.

How to prioritize remediation

1. Audit current state across all 40. Score each Strong / Adequate / Weak / Absent. Most B2B SaaS companies score Adequate or worse on 28+ of 40 at first audit.
2. Fix Tier 1 first. Entity anchor inconsistency is the single largest source of LLM hallucinations and is fixable in 30–60 days.
3. Engineer Tier 2 in parallel. G2, Capterra, TrustRadius profile completeness lifts Gemini and Copilot SOV within 30 days.
4. Sequence Tier 4 and Tier 5 over 6 months. Tier-1 press placements and earned community signals require sustained PR and community work — they cannot be sprinted.
5. Treat Tier 3, 6, 7, 8 as category-dependent. A developer-tools company prioritizes Tier 3; a Salesforce ISV prioritizes Tier 6; a hiring-tech company prioritizes Tier 7. Allocate effort accordingly.

Frequently asked questions