Is Perplexity Killing Google SEO?

Let's be precise about what's actually happening. Perplexity is not killing Google. Google still processes somewhere in the region of 8.5 billion queries a day. Perplexity processes somewhere around 10 million. This is not a one-for-one substitution — it is not even close. But the queries that are shifting aren't random ones. They're the high-intent, research-oriented, decision-stage queries that have historically driven the most valuable organic traffic. That's what makes the shift worth paying attention to now, even while the overall numbers remain small.

The more accurate framing is this: a new category of search behaviour is emerging that was always poorly served by ten blue links, and AI-native search engines are absorbing it rapidly. Understanding where that category begins and ends is the most important SEO question of 2025.

How AI Search Actually Works

Perplexity, ChatGPT Search, and Google's AI Overviews all operate on a fundamentally different model to traditional search. Traditional search ranks pages and shows you a list. AI search reads multiple pages, synthesises their content into a direct answer, and cites the sources it drew from. The end user gets an answer, not a list of candidates for an answer.

The mechanics matter for SEO. In traditional search, appearing in position one for a keyword means your page gets clicked. In AI search, appearing as a cited source means your content is surfaced within the answer — but the user may never visit your page at all. This is the zero-click problem, and it's much more acute in AI search than it ever was in traditional search.

Citation-based discovery vs. ranking

Perplexity's citation logic is not identical to Google's ranking algorithm. Google ranks based on a combination of relevance, authority, and user experience signals. Perplexity cites based on which pages directly and clearly answer the specific sub-question it is trying to resolve within a broader query. A page with lower domain authority can outperform an established publisher if it answers a specific question more precisely and in a format the AI can extract cleanly.

This is genuinely new. For fifteen years, SEO was partly a game of domain authority accumulation — build enough links to a strong enough domain and your pages rise. AI search partially decouples content quality from domain authority in a way that creates both opportunity and threat depending on your current position.

Key distinction: Perplexity doesn't crawl in real time — it pulls from its index, which is itself informed by sources like Bing. Your page still needs to be indexable and crawlable. But the ranking model that determines citation priority is different from the one that determines Google ranking, and they reward different things.

Which Queries Are Shifting — and Which Aren't

Not all search intent is migrating equally. There are types of queries where AI search is a genuinely superior experience, and types where it isn't. Understanding the split tells you where your traffic is at risk.

Queries shifting toward AI search

Comparative research: "What's the difference between X and Y?" — AI search synthesises the comparison rather than requiring the user to read three separate articles.
How-to and step-by-step: Procedural queries where the user wants a clear sequence of actions, not a list of articles about those actions.
Definition and explanation: Any query that begins with "what is," "how does," or "why does." These have always been poorly served by a list of links.
Research synthesis: Queries like "what do studies say about X" or "what's the current thinking on Y" — where the value is in aggregation, not in reading one source.

Queries staying on Google

Navigational queries: "Spotify login," "HMRC self-assessment" — people going to a specific destination.
Local intent: "Coffee shops near me," "plumber in Bristol." Google's local graph is far more capable here.
Transactional: Queries with clear purchase intent, where the user wants to compare products with prices, images, reviews — not a text answer.
News and real-time: Breaking news, live scores, current prices — anything where recency matters more than synthesis.

The pattern is clear: AI search owns the discovery and research phases of the customer journey. Google still owns the decision and action phases. For content marketers whose strategy was built around top-of-funnel informational content, this is the shift that demands a response.

GEO: Showing Up in AI Answers

Generative Engine Optimisation (GEO) is the emerging discipline of optimising content to be cited in AI-generated answers, as distinct from traditional SEO which optimises for ranking position on a results page. The two overlap significantly but are not identical.

The content that AI search cites most reliably is the content that most clearly, directly, and specifically answers a discrete question — not the content with the most backlinks or the longest word count.

What content AI search actually cites

Analysis of Perplexity citations across multiple studies in late 2024 identified consistent patterns. AI search disproportionately cites content that is:

Structurally clear: Uses headers that match the actual question being asked, not creative or vague headers. "How to set up Google Search Console" beats "Getting Started with Your SEO Journey."
Factually specific: Contains specific numbers, dates, named tools, defined processes. Vague advice ("optimise your content") is rarely cited. Specific guidance ("add your target keyword in the first 100 words of the page title and H1") gets extracted.
Authoritative but accessible: Written by identifiable authors with credentials, or published by outlets with recognised topical authority. Author bio pages and About pages matter more for AI citation than they did for traditional SEO.
Schema-marked up: FAQ schema, HowTo schema, and Article schema all help AI systems parse and classify content accurately. This is not new advice, but it has become more important, not less.

Authority signals in a zero-click world

The traditional goal of SEO was to get someone to click through to your website. GEO introduces a parallel goal: get your content cited as a source, even if the click never comes. The value in being cited is brand visibility and authority building — being the source that Perplexity users see attributed to the answer builds the kind of name recognition that drives direct search and referral traffic downstream.

This requires a reorientation in how you measure content performance. Page views are no longer the only metric that matters. Citation tracking — monitoring which AI systems reference your content and for which queries — is becoming a meaningful metric, even though the tooling to do it reliably is still early-stage.

Practical steps to show up in AI answers

Audit your existing content for question-answer structure. Every major section should begin with a clear question and answer it within the first two sentences. Don't make the AI work to find your point.
Add explicit FAQ sections to your key pages. AI search systems actively extract Q&A pairs. A well-structured FAQ at the bottom of an article is a citation target in its own right.
Implement FAQ and Article schema markup on all informational content. Validate it with Google's Rich Results Test to confirm it's being read correctly.
Build topical authority, not just keyword clusters. Perplexity and similar engines evaluate whether a site consistently covers a topic in depth, not whether it has one very long article. A cluster of ten well-linked articles on a subject outperforms one enormous pillar page.
Make your author and entity signals explicit. Named authors with linked bio pages, clear About pages, and consistent business name/address/phone (NAP) signals all improve the likelihood that AI systems identify your content as authoritative on a topic.
Don't abandon Google. AI search is growing, but Google's AI Overviews now appear on roughly 15% of queries and draw from the same pool of high-quality, well-structured content that Perplexity cites. Optimising for AI citation and optimising for Google AIO are largely the same exercise.

The honest conclusion is that Perplexity is not killing Google SEO — it is forcing it to mature. The tactics that worked when any 500-word article on a given keyword could rank are becoming less effective. The skills that have always distinguished great SEO from average SEO — deep subject matter expertise, clear writing, rigorous structure, and genuine helpfulness — are becoming the only skills that matter. That's not a threat to anyone doing SEO properly. It's a substantial threat to the content farms that have gamed keyword volume for a decade. The shift is, arguably, overdue.