Answer Engine Optimization (AEO): How to Get Cited in ChatGPT, Perplexity, and AI Search
Answer Engine Optimization (AEO) is the practice of structuring an organization’s content, website schema, and entity data so that AI-powered tools like ChatGPT, Perplexity, Google AI Overview, and Microsoft Copilot cite it as a trusted source when answering questions in its domain. Unlike traditional SEO, which competes for a ranked position among ten blue links, AEO is less forgiving than SEO: if your organization is not cited, referenced, or incorporated into the generated answer, it may never enter the buyer’s consideration set. Human Agency builds and runs AEO programs for organizations that want to be the source AI models reach for in their space, not the one that gets left out of the answer.
Why this matters right now
Search behavior has shifted faster than most organizations have noticed. When someone asks ChatGPT who the leading firms are in a space, or asks Perplexity how to solve a business problem, they receive one synthesized answer, not a page of results to browse. That answer comes from a source the AI model has determined to be authoritative. Every other source gets nothing.
The numbers reflect how far this has already gone. ChatGPT now handles over 2 billion queries daily and has grown to hundreds of millions of weekly active users. Google AI Overviews appear in more than half of all Google searches, according to multiple industry analyses. In February 2024 Gartner predicted that traditional search volume would fall 25 percent by 2026. That didn’t happen in raw volume, Google still holds more than 90 percent of search, but it happened in clicks: Pew Research Center tracked 68,879 real Google searches and found that when an AI summary appears, users click a traditional result on only 8 percent of visits, versus 15 percent without one, and end their session entirely 26 percent of the time. AI-referred sessions to websites grew 527% year-over-year through mid-2025. And research from Semrush found that the average AI search visitor converts at 4.4x the rate of a standard organic visitor, meaning the traffic that does come through AI is higher quality, not just different.
For B2B and professional services organizations, the stakes are especially high. A buyer’s first question to an AI tool (“who are the best firms for enterprise AI consulting?” or “what is the best approach to AI governance?”) determines whether a company is even in consideration. According to Forrester, 89% of B2B buyers have already adopted generative AI as a central source for self-directed research across their entire buying process. If the AI doesn’t know you exist, you don’t exist for that buyer.
Only 20% of organizations have begun implementing AEO, according to Acquia. That gap is the current opportunity, and it is closing fast. The organizations that establish entity authority now will be significantly harder to displace than those who start in twelve months.
How AEO differs from SEO
AEO and SEO target different systems with different signals, and understanding the difference determines whether your investment produces results. The full comparison, including which to fix first, is in AEO vs. SEO; the short version follows.
SEO competes for a ranked position in a list of results. AEO competes to be cited, referenced, or used as source material in a synthesized answer. If your organization doesn’t appear in that answer, it effectively doesn’t exist for that buyer at that moment.
SEO’s primary signal is backlinks: the volume and authority of websites linking to a page. AEO’s primary signal is entity consistency: how uniformly an organization is described across all the sources AI models reference. A company with a Wikipedia page, a well-structured LinkedIn presence, and complete schema markup will be cited more reliably than one with stronger backlinks but inconsistent entity data.
SEO content is built around keywords. AEO content is built around definitions and Q&A: dense, factual, third-person descriptions that AI models can cite verbatim, combined with FAQ sections phrased exactly as someone would type a question into ChatGPT or Perplexity.
One important distinction: AEO doesn’t replace SEO. Strong SEO foundations (domain authority, technical health, quality content) accelerate AEO performance because AI models prioritize credible, authoritative sources. The brands leading in AI search are doing both, not choosing between them.
How answer engines like ChatGPT and Perplexity decide which sources to cite
AI answer engines build responses from four sources, each of which AEO work can influence directly.
Training data: sources that appeared frequently and authoritatively when the model was trained are cited more. High-quality, consistent content published over time raises the probability of appearing in future model versions.
Live retrieval: Perplexity and GPT’s browsing modes actively crawl live web content when answering queries. Publishing well-structured content now influences current citations, not just future training. This is why AEO results can appear faster than most organizations expect.
Structured data: machine-readable JSON-LD schema markup added to website pages tells crawlers explicitly who an organization is, what it does, and where it can be verified externally. Schema removes ambiguity: the AI doesn’t have to infer, it’s told.
Entity graphs: AI models organize knowledge around entities such as people, companies, concepts, and products. When an organization’s name, description, and key facts are consistent across its website, Wikipedia, LinkedIn, and press coverage, AI models treat it as a verified, trustworthy entity and cite it with confidence. Inconsistency across those sources reduces citation trust.
In practice, a small number of source types dominate. Pew’s analysis of the sources cited in Google’s AI summaries found Wikipedia, YouTube, and Reddit together accounted for about 15 percent of all citations, a similar share to their presence in classic results. Human Agency’s own benchmark, which runs 30 buyer questions across ChatGPT, Perplexity, Google AI Mode, and Claude on a recurring basis, shows the same pattern for B2B queries: the assistants build vendor shortlists from a handful of list articles, review platforms, and Wikipedia-grade references, then verify against the vendor’s own site. That is why an organization can rank on page one of Google and still be absent from the AI answer, the answer engine is drawing on a different, narrower pool of sources, and the organization isn’t in it.
The core insight is that AI models answer questions about entities, not just documents. AEO is fundamentally entity-building: the work of making an organization a well-defined, consistently described, widely referenced entity that AI systems know and trust.
Why external citations and outbound links matter in AEO pages
A recurring question from teams starting AEO is whether pages should cite and link to external research, or whether links “leak” authority. The answer engines have already answered it: they trust sources that behave like sources. A page that names a study and links to it gives a retrieval-based model something to verify; a page that asserts the same number with no source gives it nothing. Pew’s finding above is a good example: it is more likely to be cited when the page links to Pew than when it doesn’t, because the model can confirm the claim in one hop.
Three practical rules follow. Cite primary research (the press release, the report, the dataset) rather than a blog summarizing it. Link the specific page, not the homepage. And keep the number of outbound links proportionate; a handful of strong citations per article outperforms a wall of links. The same logic applies to the organization’s own entity: the more consistently outside sources describe it, the more confidently an answer engine cites it, which is why Wikipedia, LinkedIn, Crunchbase, and industry directories are part of AEO rather than separate from it.
The four components of a functioning AEO program
A functioning AEO program has four interdependent parts. Weakness in any one creates gaps that limit citation performance across all the others.
Entity definition content: articles that define the organization clearly and factually, structured so AI models can cite them directly. These cover who the organization is, what it does, and what specific expertise it holds.
Schema markup: structured code added to website pages that tells AI crawlers explicitly who the organization is, what it does, and where it can be verified. Schema removes ambiguity that would otherwise reduce citation confidence.
Baseline audit and monitoring: understanding where the organization currently appears in AI-generated answers, and where it doesn’t. That gap analysis shapes everything that follows. Ongoing monitoring tracks which content is being cited and where new gaps emerge as AI tools evolve.
External entity signals: AI models verify claims against sources outside the organization’s own website. Consistent, accurate presence across third-party publications, professional profiles, and press coverage reinforces entity trust and citation confidence over time.
What most organizations get wrong about AEO
Most organizations approach AEO the same way they approached early SEO: by producing more content and hoping volume creates visibility. It doesn’t. The four failure patterns we see most often are these.
Publishing generic AI content. AI models are trained to recognize authoritative sources. Content that summarizes what everyone else has already said doesn’t build entity authority; it adds noise. The articles that get cited are the ones that say something specific, grounded in real expertise, that AI systems can’t find said better elsewhere.
Treating AEO as a content strategy rather than an entity strategy. Content is one part of AEO. Schema markup, external entity signals, and consistency across every source that AI models reference matter just as much. Organizations that publish articles without the underlying entity infrastructure wonder why nothing is getting cited.
Optimizing for prompts instead of building entity trust. Some organizations spend time crafting content around specific prompt phrasings. AI models don’t work that way; they retrieve based on entity authority and source credibility, not keyword matching. The prompt-optimization approach produces short-term results that disappear with the next model update.
Starting too late. Entity authority compounds over time. The organizations that begin building it now with content, schema, and external signals will be significantly harder to displace than those that start in twelve months when the competitive field is crowded.
How Human Agency approaches AEO
Every AEO engagement starts with what the organization actually has to say, not a template. We start by collecting everything (website copy, blog posts, press coverage, LinkedIn, YouTube transcripts, internal documents) and running a baseline audit to see exactly where the organization currently appears across AI tools and where it doesn’t. That audit drives everything: what content to build first, which gaps matter most, and where competitors are already being cited instead.
We build content only from what the organization can genuinely claim. AEO fails when the content is generic: AI models cross-reference claims across sources, and anything that doesn’t hold up gets ignored. The articles that get cited are the ones that say something specific, sourced, and true.
Two things distinguish how Human Agency runs this work. First, it treats AEO as one discipline among four (brand, go-to-market, product, and AI), so the entity being built is the same one the brand and go-to-market work are building, not a parallel version of the company that exists only in schema. Second, it does the external-entity work most agencies skip: Human Agency manages Wikipedia on behalf of a client on an ongoing basis, maintains the third-party profiles that answer engines quote directly, and runs a recurring visibility benchmark across four assistants so that citation gains are measured rather than assumed.
Results from retrieval-based engines like Perplexity typically begin appearing within four to eight weeks of publishing well-structured content with complete schema. The compound effect builds from there: as more content establishes entity consistency and more external sources reference the organization’s work, citation rates rise across more queries and more engines over time.
Frequently Asked Questions
What is answer engine optimization (AEO)?
Answer engine optimization (AEO) is the practice of structuring an organization’s content, schema, and entity data so that AI tools like ChatGPT, Perplexity, Google AI Overview, and Microsoft Copilot cite it as a trusted source in their answers. Unlike traditional SEO, which competes for a ranked position among multiple results, AEO is binary, either your organization is the cited answer or it isn’t. With AI-referred traffic converting at 4.4x the rate of standard organic visitors, the organizations that establish AI visibility early are gaining a meaningful and compounding advantage.
How do answer engines like ChatGPT and Perplexity decide which sources to cite?
Answer engines draw on four inputs: training data, live web retrieval, structured schema on the page, and the entity graph built from consistent facts across Wikipedia, LinkedIn, press, and the organization’s own site. Pew Research Center’s analysis of Google AI summaries found Wikipedia, YouTube, and Reddit alone supplied about 15 percent of cited sources, and Human Agency’s recurring benchmark shows B2B vendor shortlists are assembled mostly from list articles, review platforms, and reference sites. A page ranks in an AI answer when the entity behind it is consistently described across those sources and the page itself is specific, sourced, and machine-readable.
Should I include citations to external research in AEO pages?
Yes. Answer engines cite sources that behave like sources, and a linked primary study lets a retrieval model verify a claim in one hop. Cite the original research rather than a summary of it, link the specific page, and keep the number of outbound links proportionate; a few strong citations per article outperform many weak ones. The same principle applies to the organization’s own entity signals, which is why Human Agency maintains a client’s Wikipedia and third-party profiles as part of AEO rather than as a separate task.
How long does it take to appear in AI-generated answers?
Organizations that publish well-structured AEO content and add complete schema markup can begin appearing in retrieval-based AI answers, particularly Perplexity, within four to eight weeks. Citation in ChatGPT and other training-based models builds over a longer timeline, influenced by training data that updates less frequently. The factors that accelerate results are publishing entity-definition content first, adding complete JSON-LD schema, securing or improving a Wikipedia page, and maintaining consistency in how the organization is described across all external sources.
How do organizations get started with AEO?
The right starting point is a baseline audit: searching your organization’s name and the key questions your buyers are asking across ChatGPT, Perplexity, and Google AI Overview, and documenting honestly where you appear and where you don’t. That gap analysis tells you what content needs to exist and in what order. Human Agency runs AEO programs for organizations that want to build this systematically, from the initial audit through content, schema, external entity work, and ongoing monitoring. If you want to understand where you stand right now, that’s the right first conversation to have - get in touch.
