AI Search Optimization

Subheader

How AI is changing SEO — and what to do about it

Featured Image

close up of person's hands as they browse search results on their mobile device

Lead-in

Anyone searching the internet these days knows that an AI-powered answer holds the top spot on the page.

So it's understandable that organizations want to see their websites cited first. That's been the domain of SEO (search engine optimization) for years. But with AI, has search ranking changed?

Let's do a brief overview of what we know about AI and SEO, and what practical steps you can take to make sure your website stays visible and relevant.

Sections

The process of optimizing your website for AI search is confusing, starting with what to call it.

Some of the more popular terms you may have heard include AEO (answer engine optimization), GEO (generative engine optimization), AIO (AI optimization), and LLMO (large language model optimization). Arguably these are all slightly different disciplines, but the overall idea is the same as old-fashioned SEO — getting your content found and ranked highly. Only now we want to be found via AI, and not just Google.

With traditional SEO, this has relied on a combination of rich and relevant content, inbound links, and domain authority and trust. The older, most trusted websites with the best content were often rewarded with top rank.

With AI, this really hasn't changed too much. LLMs like Gemini, ChatGPT, and Claude are still indexing and ranking content that is widely cited, well written, and relevant. Traditional SEO remains the foundation — Ahrefs found that 76% of Google AI Overview citations in mid-2025 came from pages ranking in the organic top 10. But by early 2026, that number had dropped to 38%.

Classic SEO is necessary, but no longer sufficient.

While traditional SEO rules still apply, there are a few key differences to consider when optimizing your website for appearing in AI answers and search results.

Context matters — Unlike traditional search, AI considers the user's intent, not just their keywords. Deciphering user intent when someone types "engineering programs" on a higher-ed site is an AI superpower. It will prioritize a listing of academic programs, and not just anything with those keywords (such as an event with those words in the title).
Question splitting (or "query fan-out") — Rather than searching the exact phrase a user typed, AI tools break the question into smaller sub-queries and search for each one separately. A question like "best Drupal agency for higher ed" might become three or four separate searches behind the scenes. That means your content needs to answer the smaller questions, not just the headline one.
Broader sources of authority — Traditional SEO relied primarily on content found across websites. AI search emphasizes additional sources of expertise — traditional media, user-driven forums like Reddit and Quora, video platforms like YouTube, and reference sites like Wikipedia. Each engine sources differently: ChatGPT leans on Wikipedia and editorial sites, Perplexity on Reddit and community discussion, Google AI Overviews on a mix of YouTube, Wikipedia, and Forbes. Reddit in particular has become the single most-cited domain across major AI platforms.

When optimizing your content for AI-generated answers, there are some general guidelines to follow.

Structured content matters more than ever — Search has always benefited from well-structured content, but AI is better positioned to take advantage of it. This means both visible structure (proper HTML headings, bulleted lists) and underlying structure (Schema.org tagged elements). AI relies on these entities — people, articles, products, brands — when constructing answers.
Keep content fresh and accurate — AI systems favor recently updated, well-maintained sources. Stale content gets passed over for sites that look actively cared for.
Write to be quoted, not just read — AI tools extract snippets, not whole pages. Lead with direct, declarative answers to specific questions. Include statistics, named sources, and direct quotes — research from Princeton and IIT Delhi found that citing authoritative sources increased AI visibility by 115% for previously low-ranked content.
Build authority off your own site — Because AI weighs third-party sources heavily, your presence on Reddit, Wikipedia, YouTube, and trade publications can matter as much as your own pages. Mentions in news articles, expert roundups, and community discussions all feed the model's picture of who you are.
Make sure AI bots can actually read you — A surprising number of sites unknowingly block AI crawlers. Check your robots.txt for rules excluding GPTBot, ClaudeBot, and PerplexityBot. If you use Cloudflare, confirm its default AI bot blocking (turned on for new domains as of mid-2025) isn't quietly hiding your site from AI search. Server-side render important content too — some AI crawlers don't execute JavaScript.

After you've optimized your content for AI, how do you measure the results? How do you know how well you are doing?

Unlike traditional SEO, where Google Search Console gives you a clear view of impressions and clicks, AI search operates as a black box. But there are a few practical ways to measure your AI visibility.

Run prompt tests manually — Pick 10–20 questions a prospective customer might ask ChatGPT, Gemini, Perplexity, and Claude in your category. Run them across each platform, and track whether your brand is mentioned, cited with a link, or absent entirely. Repeat monthly to spot trends.
Set up an AI traffic channel in Analytics— By default, Google Analytics (GA4) lumps visits from ChatGPT, Claude, and Gemini into the generic "Referral" channel. You can break them out by creating a custom channel group with a regex filter for the major AI domains, then dragging that channel above "Referral" in the priority list. It takes about 15 minutes and no code.
Use a visibility tracking tool — Several platforms now automate this process, including Semrush's AI Visibility Toolkit and HubSpot's AEO Grader. They run query sets at scale, track brand mentions and citation share across engines, and benchmark you against competitors. (*note: my own recent test of HubSpot's AEO Grader was not good–but hopefully it will improve quickly or work better for others).
Watch your referrer data and direct traffic — Because so much AI-driven traffic arrives without a referrer, watch for unexplained lifts in direct visits and branded search that correlate with your content investments. A reader who saw your name in ChatGPT and typed your URL into a new tab is invisible in your AI channel, but the lift still shows up somewhere.

The regex for your GA4 AI channel

In GA4, go to Admin → Data Display → Channel Groups, and create a new group named "AI search". Create a new channel named "AI" with a source condition set to matches regex:
chatgpt\.com|chat\.openai\.com|perplexity\.ai|claude\.ai|gemini\.google\.com|copilot\.microsoft\.com|gemini-api\.google\.com|deepseek\.com|grok\.com|x\.ai
Going back to your channel list, click “reorder” and drag the new AI channel above "Referral" in the priority list before saving — GA4 matches rules top-down, so if "Referral" is listed first, your AI channel will never fire. Then check Reports → Acquisition → Traffic Acquisition and switch the channel group dropdown to your new one.
Update the regex as new AI platforms emerge. Custom channel groups apply retroactively to historical data, so you'll see prior AI visits the moment you save.

Traditional SEO optimizes content for organic search ranking. AI SEO optimizes content for high visibility across all AI-powered search results.

If that sounds like a small distinction, it isn't. The mechanics — query fan-out, third-party authority signals, citation-driven visibility — are genuinely different, and the playbook is still being written. But the foundation is the same one we've been recommending for years: clear structure, strong content, real authority, and a site that's easy for machines to read.

If you’d like to improve your site’s AI SEO, we’d be happy to chat.

About the Author

About the author