What this blog is about
For twenty years, "getting found" meant ranking a page high enough that a human would click it. That job still exists. But a second job has appeared alongside it: getting your brand named and cited *inside* the answer an AI engine writes — before anyone clicks anything. This blog is about that second job.
A few terms get thrown around for it — AI SEO, generative engine optimization (GEO), answer engine optimization (AEO). We lead with the plainest framing: AI visibility, meaning whether AI engines mention, cite and recommend you. GEO and AEO are the work you do to improve it. Everything here ladders up to one question: when a buyer in your category asks an AI a question, are you in the answer?
The one idea to hold onto
AI SEO is not a separate discipline you bolt on. Google is explicit that "AEO and GEO are still SEO" — being indexed, helpful and snippet-eligible still does most of the work. The new layer is answer-first structure (put the answer up top) and AI-crawler hygiene (let the right bots in). Treat it as SEO plus, not SEO replaced.
Why it matters now
This stopped being a "someday" topic. The usage is enormous and the click economics have already shifted under informational content.
800M
weekly active ChatGPT users (OpenAI, Oct 2025) — up from ~400M in February 2025, and that's one engine among several
58%
drop in clicks on the #1 organic result when an AI Overview appears (Ahrefs, Dec 2025) — up from 34.5% in April 2025
When an AI Overview answers the question on the page, fewer people click through — Ahrefs measured a 58% drop on the top result by December 2025. That doesn't mean SEO is dead; it means the win condition is changing. Increasingly, the prize isn't the click — it's being the brand the answer names and the source it cites. If you're not in the answer, you never enter the conversation.
A note on the scary stats
Zero-click numbers get quoted as a single settled figure; they aren't. Semrush's data actually showed zero-click queries *dropping* slightly (33.75% → 31.53%) over one window, and AI Overview trigger rates swing month to month (they peaked near 24.6% of queries in July 2025, then settled to ~15.7% by November). Treat the direction as real and the exact percentages as a range with a methodology asterisk.
Start here: the four pillars
The blog is organized around four jobs. If you only read one thing, pick the pillar that matches where you are right now.
1. Show up in AI search
The foundations: how AI engines decide which brands to name, and what gets you eligible to appear in ChatGPT, Perplexity, Gemini and Google AI Overviews. Start with How to Show Up in AI Search, then go deeper on Google AI Overviews specifically and getting AI to recommend your brand.
2. Rank in ChatGPT (and the other engines)
Each engine cites differently — they don't draw from the same pool of sources. How to Rank in ChatGPT is the pillar; from there, branch into Perplexity and Gemini & Claude. The tactics overlap, but the source preferences don't — which is exactly why a one-size answer fails.
3. Track your AI mentions
You can't improve what you can't see, and AI answers drift every time you ask. How to Track AI Mentions covers the metrics that matter — mention rate, position, sentiment, citation accuracy and share of voice — and how to run a repeatable measurement loop instead of a one-off vanity check.
4. The full AI SEO playbook
The complete reference: every tactic, in priority order, with the evidence behind it. AI SEO: The Complete Guide is the long-form pillar that ties the other three together — read it when you want the whole map in one place. For tool comparisons, see the best AI SEO tools.
Not sure where you stand? Run a free AI SEO audit — it checks whether AI engines can read and recognize your site in about 15 seconds, no signup wall.
Run a free AI SEO auditWhat actually moves the needle
Most AI-SEO advice is vibes. The highest-signal source is peer-reviewed: the GEO research paper (Aggarwal et al., KDD 2024) tested nine tactics across 10,000 queries and measured which ones actually increased visibility in generated answers. The winners are not what classic SEO trained you to do.
| Tactic | What it means | Effect on AI visibility |
|---|---|---|
| Add quotations | Quote named experts directly | Top performer (~+41% by position-adjusted word count) |
| Add statistics | Replace vague claims with specific numbers | Strong (~+31%) |
| Cite sources | Link out to the evidence behind your claims | Big lift, especially for lower-ranked pages (a #5 page saw ~+115%) |
| Keyword stuffing | The old SEO reflex | Barely registered — sometimes hurt |
The pattern is clear: write content a model can lift verbatim and trust. Beyond the lab, four operational habits do most of the real-world work:
- **Answer-first structure.** Put a concise, self-contained answer in the opening lines of each section. Engines tend to extract the first ~40–75 words — bury the answer under preamble and you hand the citation to someone else.
- **Open the door to AI crawlers.** Allow GPTBot, OAI-SearchBot, PerplexityBot and Google-Extended in robots.txt. Allowing OAI-SearchBot specifically is what lets you appear in ChatGPT search at all.
- **Earn third-party corroboration.** Reviews, directories, industry publications, Reddit and forums. Engines cross-reference outside mentions before recommending a brand — being talked about elsewhere beats hoarding on-page tweaks.
- **Signal freshness.** Perplexity in particular favors recent content; a visible current-year signal in titles tends to help citation rates.
Each engine cites a different crowd
They don't share a source pool. Directionally: ChatGPT leans heavily on Wikipedia, Perplexity leans on Reddit, and Google AI Overviews lean on YouTube — and only a small fraction of domains get cited by both ChatGPT and Perplexity. The takeaway isn't the exact percentages; it's that "one piece of content for all engines" leaves citations on the table.
Common mistakes to avoid
The fastest wins in AI SEO are usually subtraction — stop doing the thing that's quietly making you invisible. The five worth checking first:
- **Accidentally blocking AI crawlers.** A stray robots.txt rule against GPTBot, OAI-SearchBot or PerplexityBot makes you invisible to that engine regardless of content quality. Check your AI-crawler access before anything else.
- **Waiting for a magic "AI file."** There isn't one. Google states plainly that no special files, "AI text files," or schema are required to appear in AI Overviews or AI Mode. llms.txt is worth publishing as a low-cost, forward-looking bet — but it is *not* a confirmed ranking lever, and major crawlers haven't confirmed they consume it.
- **Treating GEO as separate from SEO.** Burying the answer under preamble, chasing a "GEO checklist" while neglecting indexing and helpfulness. The fundamentals still carry most of the load.
- **Over-trusting schema.** Structured data helps machines read your page, but per Google it is not required for AI features. Add it for the right reasons, not as a magic ranking input.
- **Measuring once.** AI answers are non-deterministic — ask twice, get different wording. Track visibility on a recurring schedule against real customer prompts, not just your own brand name.
Quote-accurate, on purpose
Google's exact words: "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary… You don't need to create new machine readable files, AI text files, or markup to appear in these features." We keep that straight so you don't waste a sprint chasing a file that does nothing.
Don't guess — measure
Every tactic on this blog needs a feedback loop, because AI answers shift and "it feels like we're showing up more" is not data. There are two ways to actually know.
The first is prompt testing: run a fixed set of buyer questions across ChatGPT, Perplexity, Gemini and Claude on a schedule, and track your mention rate and share of voice versus competitors over time. The second — the one most tools skip — is your own server logs: when an engine reads or cites your site, its crawler hits your pages and its answers send real referral clicks. That first-party traffic is ground truth, not a synthetic sample.
SourceWatch measures both. It tracks whether ChatGPT, Perplexity, Gemini and Claude cite your brand (visibility and share of voice against your competitor set), and it captures the real AI-crawler and AI-referral traffic hitting your site — distinguishing verified bots from spoofed ones. There's also a SourceWatch MCP server so you can pull your visibility data straight into Claude Code. Start with the free AI SEO audit to see where you stand.
See whether AI engines can read and recognize your site today.
Run a free AI SEO audit