Strategy

The Answer Engines Are Splitting Up. Here Is What That Means for Getting Cited.

Everyone is still watching Google. But ChatGPT, Perplexity, and Claude have stopped behaving like one market. Each has made a completely different bet on how it survives — and each bet changes the rules for how your brand shows up in their answers.

TL;DR

  • ChatGPT citation volumes fell an estimated 86–94% between February and April 2026 after OpenAI introduced ads for free-tier users.
  • Perplexity eliminated advertising entirely. Organic citation is the only door in — no paid placement exists.
  • Claude is staying ad-free. Content quality and structure are the whole game.
  • Two of the three largest non-Google engines now offer no paid placement at all. For those, AEO is not one tactic among several — it is the only tactic.
  • Cloudflare now blocks AI crawlers by default. Verify the bots you want can reach the pages you want cited.
Three-column graphic comparing ChatGPT, Perplexity, and Claude citation and monetization strategies

Google updated its Search Central docs on June 5 to officially name GEO and AEO as real services. Fair enough. But if you only watch Google, you are missing the more interesting story: the other answer engines have stopped behaving like one market. Here is what actually moved this spring, and what to do about it.


ChatGPT Is Citing Far Less, and That Is the Headline

The most consequential AEO development of the year is not a feature launch. It is a quiet behavioral shift inside ChatGPT.

A seoClarity analysis tracking citation patterns across five markets — US, UK, Canada, Germany, and Italy — reported that ChatGPT citation volumes fell roughly 86 to 94 percent between February and April 2026, with sharp platform-level drops around March 8 and April 19. The April change added a second lever: ChatGPT not only cited less often, it cited fewer sources when it did cite at all.

Treat the exact percentages as a single vendor’s measurement, not gospel. But the direction is not in dispute. If your AEO strategy was built on ChatGPT citations as a traffic channel, that channel is narrowing fast.

The brands that keep appearing are the ones whose content is structured to be the single source worth pulling, not one of ten links worth listing. This is exactly the kind of shift that only shows up if you are measuring it. A drop from ten citations to one looks like a catastrophe in a tracker and like nothing at all if you are guessing.


Three Engines, Three Monetization Bets

The reason ChatGPT’s behavior is changing comes into focus when you look at how each major engine has decided to make money. They diverged this year, and the strategy gap is now wide.

ChatGPT chose ads. OpenAI introduced advertising in February 2026 for Free and Go users while keeping Plus, Pro, Business, Enterprise, and Education tiers ad-free. That rollout has since expanded to the UK and to Australia, New Zealand, and Canada. Ads in an answer interface compete for the same screen real estate as organic citations, which is a reasonable explanation for why organic citations got leaner.

Perplexity killed ads entirely. After being one of the first to test sponsored answers, Perplexity wound down its advertising program, with leadership arguing that ads make users suspicious of the whole product. The consequence for brands is stark: there is no paid placement to buy. Organic citation is the only door in, and Perplexity weights freshness, clear heading structure, and original data heavily when it decides who gets cited.

Claude is staying ad-free. Anthropic has committed to keeping Claude free of advertising, positioning it as a neutral surface. No ad auction means, again, that being genuinely useful and well-structured is the entire game.

Two of the three largest non-Google engines now offer no way to pay your way in. For those, organic AEO is not one tactic among several. It is the only tactic.


The Web Is Closing to Crawlers, and Access Is Getting Priced

Underneath all of this is a fight over whether AI engines can even read your content, and on what terms.

Cloudflare, which sits in front of roughly a fifth of web traffic, now blocks AI crawlers by default and lets publishers charge them through a Pay-Per-Crawl system built on the long-dormant HTTP 402 “Payment Required” status code. Publishers get three choices per crawler: allow for free, charge per request, or block. More than 2.5 million sites now disallow AI training outright, and GPTBot is blocked by close to a fifth of sites. An early Stack Overflow pilot reportedly cut unauthorized bot traffic by about a third while lifting data-licensing revenue by roughly a quarter.

Two practical takeaways for anyone trying to be visible in AI answers:

First, robots.txt is no longer a reliable lever in either direction. Some crawlers respect it, many ignore it, and enforcement is moving to the network layer. If you want AI engines to read you, confirm you are not accidentally blocking the bots that matter — OAI-SearchBot, PerplexityBot, ClaudeBot, and friends — at the CDN or WAF level.

Second, llms.txt is a description, not a gate. It is a markdown file that tells AI models what your site is and how it is organized. It cannot restrict any crawler, and it does not replace structured data. It is a wayfinding aid, useful but frequently oversold.


What to Actually Do This Quarter

The non-Google landscape rewards the same fundamentals, harder than before.

  1. Pick three pages and build them to be the only answer. Most brands have dozens of pages that show up in AI results occasionally. That is the wrong goal right now. As citation counts shrink, occasional is the same as invisible. Choose three pages — your core category, your strongest differentiator, your most-asked question — and rewrite them with original data, a direct answer in the first paragraph, and clear H2/H3 structure. Own those three before you worry about the rest. Your AEO Audit shows you which pages are already getting cited so you know where to double down.
  2. Write to be the one source, not one of many. As engines cite fewer sources per answer, near-misses stop earning anything. Original data, clear definitions, question-led pages, and tight H2/H3 structure are what survive the squeeze. Your AEO Audit scores each of these dimensions and shows exactly where your content is falling short.
  3. Keep content fresh. Freshness is a leading citation driver, especially on Perplexity. Stale pages quietly fall out of answers. Weekly audits let you see when a page that was being cited last month has stopped appearing — before you lose the traffic entirely.
  4. Audit crawler access at the network layer. Make sure the AI bots you want can reach the pages you want cited, and decide deliberately about the ones you want to block. If you are behind Cloudflare, check that OAI-SearchBot, PerplexityBot, and ClaudeBot are on your allow list.
  5. Reallocate paid budget to content authority. For Perplexity and Claude there is no placement to buy. The money is better spent on original research, expert-authored pages, and the kind of structured content that survives a citation squeeze on every platform — including ChatGPT.

Google will keep getting the headlines. But the engines quietly rewriting the rules of citation are the ones most people are not watching closely enough. The brands that win the next year will be the ones treating each answer engine as its own channel, with its own economics, measured on its own terms.


Want to see how your brand shows up across ChatGPT, Perplexity, and Claude, engine by engine? That is what the AEO Audit is built to track.


Sources

Frequently asked questions

Common questions about AI citation strategy in 2026

The most widely reported explanation is that OpenAI introduced advertising into ChatGPT for Free and Go users in February 2026, and ads compete for the same screen real estate as organic citations. As ad inventory expanded, citation volume contracted. A seoClarity analysis tracking citation patterns across five markets documented drops of 86 to 94 percent between February and April 2026, with sharp platform-level changes around March 8 and April 19.

No. After being one of the first AI search platforms to test sponsored answers, Perplexity wound down its advertising program entirely. Leadership argued that ads make users suspicious of the whole product. The result is that there is no paid placement available in Perplexity — organic citation is the only way in, and the platform weights freshness, clear heading structure, and original data heavily when deciding who gets cited.

No. Anthropic has committed to keeping Claude free of advertising, positioning it as a neutral surface. There is no ad auction, no sponsored placement, and no paid mechanism for brand inclusion. Content quality, structural clarity, and genuine usefulness are the only levers available.

Potentially, yes. Cloudflare now blocks AI crawlers by default and has introduced a Pay-Per-Crawl system that lets publishers allow, charge, or block specific bots. If your site is behind Cloudflare, you should verify that the crawlers you want to allow — including OAI-SearchBot, PerplexityBot, and ClaudeBot — are not accidentally blocked at the CDN or WAF layer. Robots.txt is no longer a reliable lever; enforcement has moved to the network level.

Yes. The ChatGPT citation data makes this point clearly: a drop from ten citations to one looks like a catastrophe in a tracker and like nothing at all if you are guessing. Citation shifts happen at the platform level without any change on your end. The only way to know whether you are winning or losing in each engine is to measure them separately and consistently.

No credit card required

Find out where you stand.

We test the prompts your customers are typing into AI right now — engine by engine.

Get your free audit