Why AI search matters now
Google AI Overviews are shown on a majority of informational queries in 2026. ChatGPT Search and Perplexity together draw tens of millions of weekly active users. For an increasing share of B2B and B2C research, an AI answer replaces the traditional ten-blue-links result page.
The practical consequence: even if your site ranks #1 in a traditional SERP, you can be invisible in the AI answer that the user sees first.
Pages with strong external brand mentions are cited 3.2× more often than equally well-ranked pages without those mentions.
How AI picks citations
Each major AI search system follows a similar pipeline:
- 1
Retrieve
A search index (Google for AI Overviews, Bing for ChatGPT, internal indexes for Perplexity) returns candidate pages. - 2
Rerank
A language model rescore candidates for authoritativeness, freshness, brand recognition and passage relevance. - 3
Synthesize
The model writes the answer, drawing facts from top-reranked candidates and naming sources. - 4
Cite
Source citations are surfaced in the UI. The brands cited capture nearly all click-through value.
Why brand mentions matter (even without a link)
A core insight from our 2025–2026 client data: AI models pick up brand names from editorial content regardless of whether a hyperlink is present. A brand named in 50 forum threads as "the best supplier for X" gets cited by ChatGPT for that query even when most of those mentions have no link.
This is why we build brand mentions in discussions as a separate service from PR articles — both are necessary, but the un-linked mention has become a critical AI-visibility signal.
What is GEO (Generative Engine Optimization)
GEO is the discipline of optimizing for generative AI search engines. It builds on SEO and adds three new emphases:
- Brand-mention building (linked and un-linked) across authoritative sources
- Passage-level optimization — write so single paragraphs are quotable as standalone answers
- Entity-graph work — make sure Wikidata, schema.org, and major business directories all describe your brand consistently
- Source-attribution targeting — guest posts in publications AI Overviews actually cites
- Conversational keyword research — long-tail intents phrased like real questions
Practical GEO playbook
The 5-step playbook we run for every GEO-focused client:
- 1
Audit current AI visibility
Run 30–50 brand and category prompts across ChatGPT, Perplexity, AI Overviews and Claude. Document which brands are cited and which sources back them up. - 2
Identify citation sources
Find the 10–20 publications that AI models repeatedly cite for your category. Those are your priority PR targets. - 3
Build mentions and backlinks
5–10 high-quality placements per month split between editorial PR articles and authentic brand mentions in discussions. - 4
Optimize on-page citability
Restructure key pages so individual paragraphs are quotable as standalone answers. Add FAQ schema and HowTo schema where applicable. - 5
Re-measure monthly
Repeat the audit every 30 days. Track citation share, citation rank and click-through from AI surfaces.
Measuring AI visibility
The KPIs we report on monthly:
| KPI | What it measures | Healthy target |
|---|---|---|
| Citation share | % of category prompts in which the brand is cited | 15–30% for an established brand |
| Citation rank | Average position when cited (1st, 2nd, 3rd source listed) | Top 2 by month 6 |
| Source diversity | Number of distinct domains backing the citation | 5+ unique sources |
| Un-linked mentions | Brand named without hyperlink in editorial content | 20+ per month |
| AI-driven sessions (GA4) | Direct sessions tagged from ai.com, chat.openai.com, perplexity.ai | Trend up month over month |
Common mistakes
- Treating GEO as separate from SEO — they share 90% of the foundation
- Chasing one viral PR placement instead of a steady cadence
- Ignoring un-linked brand mentions as "not real" backlinks
- Forgetting entity-graph hygiene (Wikidata, Google Business Profile, schema.org Organization)
- Measuring AI visibility once and never re-measuring
Frequently asked questions
Has been running AI-visibility audits since the first AI Overviews rolled out. Writes regularly on Generative Engine Optimization (GEO) and works with international SEO agencies to build durable AI-search presence for their clients.
Sources & references
- 1Google — AI in Search — official Google announcements on AI Overviews
- 2Perplexity AI Blog — official Perplexity insights on citation pipelines
- 3OpenAI — ChatGPT Search — OpenAI announcement on web-grounded ChatGPT