AEO is the new SEO: How to get cited by AI search

Last quarter, I noticed something strange in our analytics. Traffic from "AI referrals" — visits attributed to ChatGPT, Perplexity, and Google AI Overviews — grew 340%. It's still a fraction of our organic search traffic. But the trajectory is unmistakable.

At Picsart, we serve 150 million monthly active users. A lot of them now discover our tools through AI-generated answers, not traditional search results. When someone asks ChatGPT "what's the best free background remover," we need to be in that answer. Not on page one of Google. In the answer itself.

This is Answer Engine Optimization. AEO. And it's different from SEO in ways that matter.

What Changed

Traditional SEO is about ranking. You optimize for keywords, build backlinks, improve page speed, and fight for position 1 in a list of ten blue links.

AEO is about being cited. AI models don't show ten options. They synthesize one answer from multiple sources. Your content either contributes to that answer or it doesn't exist.

The shift is subtle but profound:

Ranking ≠ Citation. Page 1 doesn't guarantee AI citation. I've seen pages ranked #1 for a keyword get completely ignored by AI Overviews because the content wasn't structured for extraction.
Comprehensiveness beats keyword density. AI models prefer content that thoroughly covers a topic over content that repeats a keyword 47 times.
Entity authority matters. AI models have internal representations of brands and entities. If your brand is strongly associated with a concept (Picsart → photo editing, Canva → design), you get cited more.
Structured content wins. AI models are better at extracting information from well-structured pages—clear headings, direct answers, comparison tables, FAQs.

The AEO Framework

I've spent the last year developing our AEO strategy at Picsart. Here's the framework we use, broken into four layers.

Layer 1: Be Crawlable

Before anything else, AI models need access to your content.

Robots.txt strategy. This is where most companies get it wrong. They block AI crawlers out of reflex—or worse, don't think about it at all.

Here's our approach: we allow AI crawlers to access our marketing pages, blog content, and tool landing pages. We block access to user-generated content and internal application pages. The logic is simple—we want AI models to know what Picsart does and recommend our tools. We don't want them training on user data.

Specific bots to consider:

GPTBot (OpenAI) — powers ChatGPT's browsing and training
anthropic-ai and ClaudeBot (Anthropic) — powers Claude
PerplexityBot — powers Perplexity search
Google-Extended — controls Gemini training (separate from Googlebot)

Each has a different impact. Blocking GPTBot means ChatGPT can't browse your site for real-time answers. Blocking Google-Extended affects Gemini training but not your Google Search ranking.

My recommendation: allow all AI crawlers on your public marketing content. The visibility upside outweighs the training data concern for most B2C products.

Sitemap coverage. Make sure your AI-optimized pages are in your sitemap. AI crawlers follow sitemaps just like Googlebot. If a page isn't in your sitemap, it might not get crawled.

Layer 2: Structure for Extraction

AI models extract information from your pages. The easier you make extraction, the more likely you get cited.

Direct answers in the first paragraph. When someone asks "what is a background remover," the AI model scans your page for a direct answer. If your first paragraph is a fluffy introduction about the history of photo editing, you lose.

We restructured our landing pages to lead with a direct, one-sentence answer to the primary query. Then we expand with context, features, and use cases.

Before:

"In today's digital world, removing backgrounds from images has become an essential skill for photographers, designers, and social media creators alike."

After:

"A background remover is a tool that automatically detects and removes the background from a photo, leaving just the subject on a transparent canvas."

The second version is what AI models cite. It's direct, factual, and self-contained.

Comparison tables. AI models love structured comparisons. When someone asks "Picsart vs Canva for background removal," a well-formatted comparison table on your page gives the AI exactly what it needs.

We add comparison sections to our key landing pages. Not as competitive takedowns—as honest, factual comparisons that position our tool clearly.

FAQ sections with schema markup. FAQPage schema is one of the highest-signal structured data types for AEO. Each question-answer pair is a discrete, extractable unit that AI models can cite directly.

We generate FAQ sections programmatically based on actual search queries from Google Search Console. Real questions from real users, answered directly.

How-to content with step markup. HowTo schema gives AI models a clear sequence of steps to cite. Our tutorial pages use this extensively.

Layer 3: Build Entity Authority

This is the long game. AI models maintain internal representations of entities—brands, people, concepts. The stronger your entity associations, the more you get cited in relevant queries.

Consistent entity descriptions. Everywhere Picsart appears on the web—our site, Wikipedia, review sites, press coverage—the description should be consistent. "AI-powered photo and video editing platform" or similar. This reinforces the entity association in training data.

Topical authority through content depth. AI models don't just look at one page. They assess whether your domain is authoritative on a topic based on the breadth and depth of your content.

We publish extensive content around our core tools: background removal, AI image generation, photo enhancement, video editing. Not thin, keyword-stuffed pages—genuinely useful guides, tutorials, and comparisons. This builds the topical authority that makes AI models trust our content.

Third-party mentions. AI models weight information from multiple independent sources. Press coverage, review sites, industry publications, and comparison articles that mention your brand all contribute to entity authority.

This isn't new—it's basically digital PR. But the mechanism changed. Instead of building backlinks for PageRank, you're building mentions for AI model confidence.

Layer 4: Monitor and Iterate

AEO is measurable, but the tools are different from traditional SEO.

AI referral tracking. Set up proper attribution for traffic from AI sources. In Google Analytics, create segments for referrals from chat.openai.com, perplexity.ai, and other AI platforms. Track this over time.

Citation monitoring. Regularly query AI models with your target keywords and check if you're cited. We do this weekly for our top 50 keywords. "What's the best free background remover?" "How to remove image background?" "Best AI photo editor?"

It's manual and tedious. The tooling will get better. But for now, you need to do the work.

AI Overview presence. Google AI Overviews show which pages contribute to AI-generated answers in Google Search. Track which of your pages appear in AI Overviews for your target queries.

We use a combination of Ahrefs (which now tracks AI Overview presence) and manual spot-checks to monitor this.

Tactical Playbook

Here's the specific checklist we follow when optimizing a page for AEO.

Content structure:

Lead with a direct answer to the primary query (first 1-2 sentences)
Use clear H2/H3 hierarchy that matches question patterns
Include a comparison table if the query implies alternatives
Add an FAQ section with 5-8 real user questions
Use numbered lists for processes and how-tos

Schema markup:

FAQPage for FAQ sections
HowTo for tutorial/process content
Product for tool pages
Article for blog content
Organization for brand pages

Robots.txt:

Allow GPTBot, ClaudeBot, PerplexityBot on marketing content
Block on user-generated content and app pages
Review quarterly as new bots emerge

Content quality signals:

Author attribution with real names and credentials
Published and updated dates visible on page
Sources and references for claims
Original data or research when possible

AEO vs. SEO: They're Not Competitors

I want to be clear: AEO doesn't replace SEO. It's an additional layer.

Google still sends the vast majority of our traffic. Backlinks still matter. Page speed still matters. Technical SEO still matters. Everything I wrote about in automating SEO with AI and LP conversion optimization still applies.

But the marginal return on AEO investment is enormous right now because almost nobody is doing it systematically. Most companies are still debating whether to block AI crawlers. Meanwhile, the companies that optimize for AI citation are building an early-mover advantage that compounds.

The parallel to early SEO is striking. In 2005, most companies didn't have an SEO strategy. The ones that started early dominated for years. AEO is at that same inflection point.

What's Coming

A few predictions based on where I see things heading:

AI referral traffic will 10x in the next 18 months. As ChatGPT, Perplexity, and Google AI Overviews become default search behaviors, the traffic from AI citations will grow dramatically. We're seeing the early hockey stick.

Citation attribution will improve. Right now, AI models are inconsistent about citing sources. This will get better as publishers push back and platforms compete on trust. Better attribution means better measurement, which means more investment in AEO.

Structured data becomes critical. As AI models get better at parsing schema markup, the ROI on implementing comprehensive structured data will increase. Companies with clean schema across their sites will have a significant advantage.

Entity authority becomes a moat. Building strong entity associations in AI models takes time and consistency. Companies that start now will be hard to displace later—similar to how domain authority in traditional SEO compounds over years.

Getting Started

If you're starting from zero, here's my priority order:

Audit your robots.txt. Make sure you're not accidentally blocking AI crawlers on your marketing content.
Pick your top 10 pages by traffic. Restructure them with direct answers, FAQ sections, and comparison tables.
Implement FAQPage schema on those pages. This is the highest-impact schema type for AEO.
Set up AI referral tracking in your analytics. You can't improve what you don't measure.
Start monitoring AI citations weekly for your top keywords.

The beauty of AEO is that most of the tactics also improve traditional SEO. Better content structure, comprehensive coverage, schema markup—these help with Google rankings too. You're not choosing between SEO and AEO. You're doing both.

The companies that figure this out early will own the AI search landscape. The ones that wait will wonder why their organic traffic is declining despite ranking well.

The answer engines are coming. Make sure they know who you are.