AI Search Visibility in 2026: Ranking on ChatGPT, Perplexity, Gemini, and Claude
Short answer: to rank in ChatGPT and Perplexity, let their search crawlers in, keep the page indexed, and lead with a direct, sourced answer. OpenAI says OAI-SearchBot surfaces sites in ChatGPT search and recommends allowing it in robots.txt; Perplexity says the same of PerplexityBot. Getting cited in those answers (Generative Engine Optimization, or GEO) overlaps with classic SEO but rewards source-rich, specific content with clear entity signals.
Our data: the CPM, CPC and corporate-domain figures on our case studies page come from real Dupple campaigns; we publish no AI-assistant citation-share numbers here.
What changed: Updated October 2026: statistics and engine claims re-checked against vendor documentation, and every figure that could not be traced to a first-party source was removed.
Where AI search share sits in 2026
Vendor documentation is the safe ground here, so this guide sticks to it:
- ChatGPT Search: OpenAI's own crawler documentation describes OAI-SearchBot as the bot that surfaces websites in ChatGPT's search features
- Perplexity: PerplexityBot surfaces and links websites in Perplexity search results
- Google AI Overviews and AI Mode: Google's AI features documentation says supporting links come from pages that are indexed and eligible to show in Google Search with a snippet
- Claude.ai: web search launched in March 2025
We do not publish usage-share numbers for these assistants because none of the vendors above publishes a comparable figure on those pages. Classic blue-link SEO still matters but increasingly feeds the AI answer rather than driving direct clicks.
How the AI engines actually pick sources
Different models, different ranking logic. Oversimplified but useful:
ChatGPT Search
- Uses its own search crawler, OAI-SearchBot: sites opted out of it will not be shown in ChatGPT search answers, though they can still appear as navigational links (OpenAI)
- OpenAI says a robots.txt change can take about 24 hours to take effect for search results
- OpenAI does not publish its ranking factors, so freshness, schema and answer extractability are practical guidance, not documented weights
Perplexity
- PerplexityBot surfaces and links websites in Perplexity search results; Perplexity recommends allowing it in robots.txt and its published IP ranges (Perplexity)
- Perplexity does not document its ranking factors on that page; content specificity and sourcing are practical guidance
Google AI Overviews / Gemini
- Links come from pages that are indexed and eligible to appear in Google Search with a snippet
- Google says there are no additional requirements for AI Overviews or AI Mode, no special optimizations and no special schema.org structured data (Google Search Central)
- Google notes these features may use a "query fan-out" technique, issuing several related searches while building an answer
Claude.ai (web search)
- Launched web search in March 2025, available on all Claude plans since May 2025
- Anthropic's own post says web search became available globally on all Claude plans on May 27, 2025
- Anthropic does not document source-ranking factors there, so depth and primary sources are practical guidance
What gets cited in AI answers
We removed the citation-correlation table that used to sit here: it could not be traced to a published study we can link. What the vendor documents above support is narrower: crawl access, indexing and a clear, extractable answer. The rest of the playbook below is practical guidance from how these engines present sources, not measured lifts.
The signal is: AI engines want content that would pass as a research briefing. Specific numbers, sources, clear structure, and a direct answer near the top.
The GEO playbook that works in 2026
Start every article with a direct answer
The TL;DR paragraph matters more for AI visibility than for classic SEO. The first lines are the easiest part for an assistant to quote. Write it like a research abstract.
Bad opening: "Newsletter sponsorships are an exciting way for B2B marketers to grow..."
Good opening: "Newsletter advertising for B2B tech companies is sold per placement: a Primary Ad on a 800,000-subscriber list is the first sponsored placement at the top of the email, at a flat rate in the media kit."
Use numeric specificity
Specific numbers are easier to quote than ranges or adjectives. "Fast" becomes "under 250ms". "Affordable" becomes "from $29/month". "Many marketers" becomes "38% of B2B marketers in 2025."
Every claim you can quantify, quantify.
Add structured data
JSON-LD schema types worth having:
- Article (author, datePublished, dateModified, publisher)
- FAQPage (question/answer pairs)
- HowTo (step-by-step content)
- Product / SoftwareApplication (for tool pages)
- Organization (on your site-wide footer or about page)
Schema doesn't directly rank you. Google states no special structured data is needed for AI Overviews, so treat schema as clean extraction, not a ranking lever.
Cite primary sources
Readers and answer engines both need to see where a number comes from. Link 3-5 primary sources (original studies, authoritative docs, direct data) and say when you read them.
This runs contrary to old-school SEO advice about "hoarding link equity." In 2026, citations strengthen your position.
Include original data or proprietary insight
The single strongest predictor of being cited: the page contains data or insight the AI engine can't get elsewhere. Original survey data, benchmark studies, case study metrics, expert opinions, anything that makes your page the source.
Our case studies are an example: the specific CPM, CPC, and corporate-domain numbers from real campaigns are the kind of data an answer about "newsletter advertising performance" can only get from us.
Name entities explicitly
Pages that explicitly name products, people, companies, and concepts are easier to match to a question. "The CRM" is weaker than "Salesforce." "The observability tool" is weaker than "Datadog." Name things.
Keep content fresh
Fresh content tends to suit queries with temporal intent, especially for queries with temporal intent ("in 2026", "current", "latest"). Set a dateModified in your schema and actually update the content, a simple timestamp change without real updates helps nobody.
Build a topic cluster, not isolated pages
AI engines look at site-level topical authority. A site with 20 interlinked articles on B2B lead generation gives an engine more context on who you are than one standalone article.
This is where our content hub strategy works: the newsletter advertising rates page, CPM benchmarks, sponsorship ROI, and lead gen playbook cross-link and reinforce each other.
What's different from classic SEO
| Factor | Classic SEO | AI Search |
|---|---|---|
| Backlinks | Very high weight | Medium weight |
| Keyword density | Medium | Low (can hurt) |
| Domain age | Medium | Low |
| Direct answer at top | Medium | Very high |
| Numeric specificity | Low | High |
| Primary source citations | Low | High |
| Schema markup | Medium | Not required (Google) |
| Original research/data | Medium | Very high |
| Site topical depth | Medium | Very high |
| Mobile performance | High | Medium |
| Click-through rate signals | High | Low (AI often doesn't drive clicks) |
These ratings are our editorial judgment, not measured weights, and none of the vendors publishes ranking weights. The convergence: content that's genuinely useful, specific, and well-structured wins in both. The divergence: AI de-emphasizes link-building games and over-weights source credibility + extractability.
How to measure AI visibility
Tools that currently track AI citations:
- Ahrefs Brand Radar: runs prompts through AI Overviews, AI Mode, ChatGPT, Perplexity, Gemini and Copilot; custom prompt tracking starts at $50/month and the AI Visibility Index at $199/month (read on Ahrefs, October 2026)
- Semrush AI Overviews tracking: citation tracking in Google's AI snippets
- Profound / Athena / Peec AI: purpose-built GEO analytics
- Manual sampling: test 20-30 key queries monthly across ChatGPT, Perplexity, Gemini, Claude
Metrics to track:
- Citation rate: % of relevant queries where your site appears as a source
- Share of voice: your citations vs. competitor citations on the same queries
- Answer position: first citation vs. later citations in the answer
- Click-through from AI: analytics shows the referrer as
chatgpt.com,perplexity.ai,gemini.google.com
How do you track AI search visibility in practice?
Our earlier benchmark ranges for citation rates and share-of-voice leaders were removed because we could not link a source for them. Build your own baseline instead: run the same 20-30 queries every month, log which domains are cited, and compare against competitors. The pattern worth testing is whether the assistants cite fewer sources per answer than Google shows links.
Improve the sources AI already trusts
Sponsored Measurement shows where your brand is missing, but visibility also depends on the third-party pages that AI engines cite. Noble helps teams address that off-site layer by finding relevant cited articles, coordinating accurate new mentions, and refreshing outdated or unfair references. This complements on-site improvements: instead of only restructuring your own pages, you can work toward appearing in the external sources buyers and AI systems already use.
The tactical 90-day GEO plan
Weeks 1-2: Audit
- Identify your 10-20 most important search queries
- Test them in ChatGPT, Perplexity, Gemini, Claude
- Record: which competitors are cited, what content format wins, what's your current position
Weeks 3-4: Foundation
- Allow OAI-SearchBot and PerplexityBot in robots.txt
- Add JSON-LD schema to all content pages
- Rewrite article openings to include direct answers
- Add FAQ sections to top pages
- Ensure
dateModifiedis accurate and content is current
Weeks 5-8: Content
- Publish 3-5 deep content pieces on core topics (2,000+ words each, original data)
- Interlink existing content to form a topic cluster
- Add author bylines with credentials
Weeks 9-12: Distribution + tracking
- Promote content for real citations (newsletters, communities, earned media)
- Set up AI citation tracking (Profound, Ahrefs Brand Radar, or manual)
- Measure baseline citation rate and share of voice
- Iterate on top-performing patterns
The biggest mistake: treating AI search as a traffic channel
AI search often doesn't drive clicks, and classic search is heading the same way: zero-click search ended 58.5% of US Google searches without a visit to any site in 2024 (SparkToro). The value is brand visibility and mind-share, the user sees your brand cited as the source of a useful answer. That compounds into category authority, branded search, and inbound.
If you measure AI visibility only by referrer traffic, you'll undervalue it. Measure citation rate and share-of-voice first; click-through is secondary.
FAQ
How do you get cited by ChatGPT and Perplexity?
Allow OAI-SearchBot and PerplexityBot in robots.txt, as OpenAI and Perplexity recommend in their documentation. Then make the page worth quoting: a direct answer in the first lines, dated figures, and links to primary sources.
How does ChatGPT choose sources?
OpenAI documents that OAI-SearchBot surfaces sites in ChatGPT search and that opted-out sites are not shown in answers. It does not publish ranking factors, so the rest is practical guidance.
What is GEO vs SEO?
SEO earns a ranking in a list of links. GEO earns a citation inside a generated answer. The inputs overlap (indexing, useful content), so most pages need both.
How do you track AI search visibility?
Use a tracker such as Ahrefs Brand Radar (custom prompt tracking starts at $50/month, read October 2026) or sample 20-30 key queries monthly and log which domains are cited.
Related reading
- Why your brand is not showing up in AI search
- Newsletter advertising rates
- B2B tech lead generation
- Best lead gen channels for B2B tech
For a backlink in an AI-cited article that compounds your AI visibility, Dupple's backlink placements (pricing on request) are indexed by Google, Bing, and AI engines.
Some offers on this page may be paid placements or contain affiliate links.
Across our whole catalogue the median is $23.50 a month and 37% of tools advertise a free tier. Figures are the lowest recurring monthly price quoted in our own review of each tool, read on 2026-10-23. Full method and every category.
