For years, the SEO playbook was straightforward: earn backlinks, climb rankings, capture clicks. But as AI reshapes how traditional SEO works, a different mechanism is determining which content gets seen — and it’s not backlinks. It’s citations. The role of citations in AEO is fundamentally different from link-building: instead of other publishers vouching for your page, AI answer engines are selecting your content as the direct source behind their generated answers.

This shift matters because the stakes are tangible. When ChatGPT, Perplexity, or Google’s AI Overviews cite your page, that’s not a ranking boost in a list of blue links. It’s your content becoming the answer for a growing share of buyers who never scroll to traditional results. And with AEO tools and best practices now available to measure and optimize this visibility, citations in AEO are no longer theoretical. It’s trackable, improvable, and directly tied to the pipeline.

In this guide, I’ll break down exactly how AI engines select citations, what type of content earns them, and how to build a citation strategy that drives measurable AI visibility using generative engine optimization tools and HubSpot’s integrated platform.

Table of Contents:

Why Citations Matter for Answer Engine Optimization (AEO)

First, let me be direct: Citations aren’t the entire point of winning AEO. They are, however, one of the clearest signals that your content is working inside the systems that now shape how buyers find answers.

The search landscape has fundamentally shifted. According to HubSpot’s 2026 State of Marketing Report, 49% of marketers agree that web traffic from search has decreased due to AI-generated answers. Yet, 58% note that AI referral traffic carries much higher intent than traditional search.

As an Associate Content Writer and Marketer at HubSpot, I’ve witnessed this firsthand: while blog traffic has declined, leads from LLMs are up 1,850% and convert 3x better. That conversion gap is why citations deserve serious attention from every marketing team allocating resources right now.

Meanwhile, 42% of CRM software buyers now use AI search as part of their evaluation process. When nearly half your potential buyers are asking ChatGPT or Perplexity instead of Google, being cited in those AI-generated answers becomes a direct pipeline driver rather than a vanity metric.

What do citations actually do in AEO?

AI answer engines select citations based on:

• Clarity
• Authority
• Structure
• Content freshness

When an LLM like ChatGPT, Gemini, or Perplexity generates a response, it draws on sources it considers trustworthy, well-structured, and semantically clear. A citation in that context means your content was the answer… or part of it.

The role of citations in AEO becomes clearer when you compare how AI engines evaluate content versus how traditional search engines do:

• Backlinks in SEO signal domain authority through link volume, anchor text, and the quality of referring domains. They tell Google, “Other sites vouch for this page.”
• Citations in AEO signal source reliability through content structure, factual density, and semantic clarity. They tell an LLM, “This content directly and accurately answers the user’s question.”

Both matter. But 41% of marketers say updating their SEO strategy for search changes is the top trend they’re exploring. The distinction is critical: You can have strong backlinks and still never appear in an AI-generated answer if your content isn’t structured for machine readability.

Citations are only one metric in the AEO era.

A complete picture of AEO success includes multiple signals beyond citation counts:

• AI visibility score: How frequently and prominently your brand or content surfaces in AI-generated responses. (Tools like HubSpot’s AEO Grader let you benchmark this directly.)
• LLM referral traffic: The volume and quality of visitors arriving from AI platforms (trackable in Marketing Hub alongside your traditional organic channels).
• Conversion rate from AI referrals: As HubSpot’s own data shows, these visitors convert at significantly higher rates, making this a revenue-tier metric.
• Brand mention frequency: Whether AI engines reference your brand by name, even without a clickable link.
• Answer inclusion rate: How often your content appears in synthesized AI answers for your target queries.

Citations serve as a proof point that your content strategy aligns with how AI engines discover, process, and surface information.

How AI Engines Select Citations and Sources

AI answer engines select citations based on:

• Clarity
• Authority
• Structure
• Freshness of content (not on backlink volume)

Understanding this distinction is the single most important shift for teams moving from traditional SEO to an AEO-first strategy.

When a user asks ChatGPT, Claude, or Perplexity a question, the underlying process differs significantly from how Google ranks a list of blue links.

To help you understand how AI engines decide what to cite, I’ve broken down exactly what actually happens when an AI engine generates an answer and assigns sources:

• Retrieval. The AI engine queries an index (or the live web, in Perplexity’s case) to pull a set of candidate sources that match the user’s intent. Content that uses clear headings, direct definitions, and structured data is more likely to surface during this step.
• Evaluation. The model assesses each candidate for factual density, source authority, semantic clarity, and how directly the content answers the query. Vague, keyword-stuffed pages get filtered out, even if they have thousands of backlinks.
• Synthesis. The engine combines information from its top-evaluated sources into a single generated response and attributes citations to the specific pages it drew from.
• Citation assignment. Not every source used during synthesis earns a visible citation. The model selects the sources that contributed the most direct, verifiable claims to the final answer.

Each AI agent type handles this process slightly differently:

• Perplexity cites inline with numbered references on every response.
• ChatGPT (with browsing enabled) links to sources selectively.
• Google’s AI Overviews pull from indexed pages and feature expandable source cards.

But across all of them, the underlying selection criteria converge on the same core signals. The five signals AI engines weigh most heavily when selecting citations are:

• Topical authority and depth. Does this source demonstrate comprehensive expertise on the subject, or is it a surface-level overview? Pages that cover a topic with rich factual detail, supporting data, and clear entity relationships get cited more often.
• Structural clarity. Content organized with descriptive H2s/H3s, definition-style opening sentences, and logical hierarchy is easier for models to parse and quote accurately.
• Factual specificity. AI engines prefer content that states verifiable claims (statistics, named frameworks, dated research) over content that hedges with phrases like “some experts say” or “it’s generally believed.”
• Freshness. Regular content updates help signal freshness to AI citation systems.
• Source reputation. Domain-level trust still matters, but it’s evaluated differently than Domain Authority in SEO. AI engines weigh whether a source is consistently accurate, frequently referenced across the web, and recognized within its subject area.

Pro Tip: You don’t need to guess which of these signals your content is hitting or missing. Run your priority pages through HubSpot’s AEO Grader to benchmark your AI visibility and identify specific structural or content gaps that may be costing you citations.

Citation Types and What They Prioritize

Citations become especially clear when you compare what earns visibility across different engines.

Below, I’ve created a chart that categorizes citations by type, AI engine, and what each citation style prioritizes. Take a look:

However, structured data and schema markup increase the likelihood of being cited by AI. If your pages lack the following, you’re making it harder for AI engines to confidently extract and attribute your content, even if the written content itself is excellent:

• FAQ schema
• HowTo schema
• Article structured data

This is a best practice for AI search visibility that carries over directly from SGE optimization into broader AEO work.

Overall, citations within AEO ultimately come down to this: AI engines aren’t counting who links to you. They’re evaluating whether your content is the most clear, structured, authoritative, and current answer to the question being asked.

Pro Tip: Teams that internalize this shift and track it through tools like HubSpot AEO will capture the high-intent AI referral traffic that’s already reshaping how buyers discover solutions.

The Role of Citations in AEO

The way people find information is splitting in two, and citations are the connective tissue between your content and AI-generated answers.

Understanding citations in AEO starts with understanding just how fast this shift is happening, and why the old playbook of chasing backlinks alone no longer covers the full picture.

Here’s what you need to know:

1. AI search adoption is accelerating faster than most teams realize.

The numbers paint a clear trajectory. Gartner projects that traditional search engine volume will drop by 25% by 2026, as search marketing loses market share to AI chatbots and virtual agents. That’s not a distant forecast. It’s happening right now, right before our eyes.

On the consumer side, adoption is already mainstream. Here’s the data to prove it:

• 34% of U.S. adults said they had used ChatGPT as of June 2025, roughly double the figure from 2023, according to Pew Research Center.
• As shared by Stan Ventures, Google’s AI Overviews reached over 1.5 billion users per month in Q1 2025 (that’s 26.6% of all internet users worldwide).
• AIOs now appear for 9.46% of all keywords on desktop (16% in the U.S.) and 12.8% or more of all Google searches by volume, according to Amsive research.

2. Citations are your content’s entry point into AI answers.

In traditional SEO, backlinks function as votes of confidence. Other sites linking to yours signal authority to Google’s ranking algorithm.

Citations in AEO work differently. They are direct attributions: An AI engine selecting your content as the source behind a specific claim in a generated answer.

Citations in AEO differ from backlinks in SEO in several important ways:

• Backlinks are created by other publishers linking to your page. You earn them through outreach, PR, and content quality over time. They influence rank position in a list of results.
• Citations are assigned by AI models during answer generation. You earn them through structural clarity, factual specificity, and topical authority. They influence whether your content is the answer.

AEO citations matter because when ChatGPT, Perplexity, or Google’s AIO cites your page, the answer engine is telling the user: “This is where this information comes from.” That’s a trust signal with direct downstream impact on brand visibility, referral traffic, and conversion.

Pro Tip: Use HubSpot’s AEO Grader to check whether your priority pages are currently being cited (or even surfaced) in AI-generated answers. Many teams assume their top-ranking SEO pages also perform well in AEO. They’re often not.

3. AI Overviews (AIOs) are reshaping click behavior, and citations are the new click-drivers.

2026 Amsive data reveals a nuanced picture of how AIOs are changing search behavior:

• AIOs are reducing clicks by 34.5% on queries where they appear.
• They show up disproportionately for informational queries, longer search queries, and queries with higher search volumes (exactly the kind of top-of-funnel content most marketing teams invest heavily in).
• They appear less frequently for branded and local queries, as well as for shorter search queries.
• They predominantly surface on non-monetized searches, meaning the queries they’re reshaping are the informational ones people weren’t bidding on anyway.

Here’s why this matters specifically for citations: “When an AIO lowers clicks to regular search results, the sources it cites are most likely to get the remaining clicks.”

Citation concentration — the degree to which a small number of sources dominate AI-generated citations — is high (according to 2026 research from an Ahrefs study, the top 50 domains account for 28.90% of all AIO mentions). If your content earns a citation in an AIO, you’re capturing visibility that would otherwise be lost entirely.

4. The role of citations in AEO is measurable, not theoretical.

One of the biggest barriers teams face is the perception that AEO is vague or unmeasurable. However, I’d like to propose a different, perhaps controversial argument: It’s not.

AEO citations connect directly to trackable outcomes, such as:

• Citation presence: Is your content appearing as a source in AI-generated answers? HubSpot’s AEO Grader measures this directly against your target queries.
• LLM referral traffic: Marketing Hub lets you segment traffic arriving from AI platforms separately from organic search, so you can see exactly how much pipeline AI visibility is driving.
• Click-through from citations: When your page is cited in a Perplexity answer or Google AIO, you can track the resulting visits and conversions just like any other referral channel.
• Brand mention frequency: Even when citations don’t include a clickable link, brand mentions in AI answers build recognition and trust that influences downstream search and direct traffic.

5. Freshness and depth determine citation durability.

Earning a citation once ain’t the same as keeping it. Regular content updates support freshness signals for AI citations, meaning stale content is replaced by competitors who publish more current data, frameworks, or examples.

AI engines re-evaluate sources continuously. A page that was cited in March may lose that citation by June if a competitor publishes a more current, more comprehensive version of the same answer. (This is especially true for data-driven content, industry benchmarks, and anything tied to evolving best practices, which describes most B2B marketing content.)

Citations in AEO depends on maintaining two things over time:

• Depth: Content that covers a topic comprehensively, with specific data points, named frameworks, and clear entity relationships, earns citations more consistently than surface-level overviews.
• Freshness: Scheduling and content audit tools (like HubSpot’s Content Hub) let teams systematize update cycles so that high-priority pages stay current without relying on manual memory.

This is where AEO diverges most clearly from traditional SEO maintenance. In SEO, a well-linked evergreen page can hold its ranking for years with minimal updates. Conversely, in AEO, FAQs and knowledge graphs help AI engines extract and cite accurate information, provided that the information reflects the current reality.

That said, outdated statistics, deprecated tools, and old screenshots are citation killers.

Citation is the mechanism, visibility is the outcome.

The reason citations in AEO deserve dedicated strategic attention comes down to a simple pipeline reality: A quarter of internet users already interact with AI-generated answers monthly.

With traditional search volume declining, the AI answers replacing clicks reward a fundamentally different set of content attributes than those most SEO programs were built around.

Citations are the mechanism through which your content earns visibility in this new layer of search. But they’re not the only AEO metric that matters. These other signals carry weight, too:

• Brand mentions
• AI referral conversion rates
• Answer inclusion rates

These metrics all contribute to the full picture. But citations are the most tangible proof point that your content is structured, authoritative, and current enough to be selected as an AI engine’s source of truth.

What type of content gets cited the most in LLMs?

Here’s the thing: Hyper-specific content that demonstrates true expertise gets visibility across LLMs. Generic, AI-generated fluff won’t achieve meaningful visibility in the new search ecosystem, and the data backs this up clearly.

You see, we’re entering a period in which the bar for “good enough” content has risen. When AI engines can generate passable surface-level answers on their own, they don’t need to cite your page for restating what they already know.

They cite sources that add something they can’t generate independently, which happens to be:

• Original data
• Specific frameworks
• Named methodologies (like Loop Marketing, wink wink)
• Expert analysis grounded in real experience

Citations reward depth, not volume.

1. Earned content dominates AI citations; owned content alone isn’t enough.

2026 research from Search Engine Journal reveals a finding that should reshape how teams think about content strategy: across all AI platforms, earned content accounts for the largest share of citations, while user-generated content (UGC) is increasingly represented. (TLDR — “earned content” is content about your brand that other people create — press mentions, reviews, third-party coverage, and organic social posts you didn’t pay for or publish yourself..)

This means the content most likely to be cited by AI engines isn’t just what you publish on your own domain. More specifically, it’s:

• Coverage
• Mentions
• Reviews
• Discussions happening about your brand on third-party sites

Thus, the implication for citations in AEO is significant:

• Earned content (press coverage, industry publications, expert roundups, third-party reviews) gets cited most frequently across LLMs.
• UGC (forum discussions, community posts, user reviews) is growing as a citation source. AI engines increasingly treat authentic user perspectives as valuable reference material.
• Owned content (your blog, your resource center, your landing pages) still matters, but it’s not sufficient on its own.

Pro Tip: Earning mentions on trusted third-party sites may be even more valuable than optimizing your domain content alone. Invest in a mix of owned content, third-party coverage, and presence on relevant UGC platforms to increase the likelihood of being cited by AI search engines. Then, track which third-party mentions are driving AI visibility alongside your owned content performance in Marketing Hub.

2. You don’t need to be a top-tier domain to earn citations.

One of the most encouraging findings from Search Engine Journal’s quality distribution analysis is that AI engines cite across a wide quality spectrum — not just from elite publishers:

• High-quality sources: ~31.5% of citations
• Upper-mid quality sources: ~15.3% of citations
• Mid-quality sources: ~26.3% of citations
• Lower-mid quality sources: ~22.1% of citations
• Low-quality sources: ~4.8% of citations

The big takeaway here? AI engines prefer higher-quality sources, but they often cite middle-tier sources when those sources provide the clearest, most specific answers.

So, here’s what this means for your team: If you’re not the New York Times or Harvard Business Review, you can still earn citations by producing content that is more specific, better structured, and more factually dense than what larger competitors publish on the same topic.

3. The content attributes that earn citations vs. the ones that don’t.

Based on citation quality distributions and earned content data, a clear pattern emerges about the types of content LLMs actually select as sources.

Here’s what separates content that earns AI citations from content that gets ignored:

Building a Citation-Earning Content Strategy

Citations within AEO depend on a deliberate strategy that spans owned, earned, and community-driven content.

After all the data I’ve shared within this post thus far, here’s what to decipher from it and prioritize in your evolving AEO content strategy:

• Lead with original insight. Every piece of content should contain at least one data point, framework, or perspective that doesn’t exist anywhere else on the web. This is the single strongest citation driver.
• Invest in earned coverage. PR, guest contributions to industry publications, participation in expert roundups, and podcast appearances all create third-party content that AI engines can cite, often more readily than your owned pages.
• Show up where UGC happens. Community forums, LinkedIn discussions, Reddit threads, and review platforms are increasingly cited by LLMs. Having your brand or team members present in these spaces (contributing value, not just promoting) builds the kind of distributed authority that AI engines reward.
• Structure for extraction. Use Content Hub to implement schema markup, clear heading hierarchies, and definition-style lead sentences that make it easy for AI engines to identify and attribute your claims.

AEO citations ultimately come down to whether your content adds to the knowledge landscape or just restates it. AI engines have access to the sum of published information; they cite sources that contribute something distinct.

The teams that internalize this standard and build it into their editorial workflow will consistently earn citations, while those producing interchangeable content will remain invisible, regardless of how many backlinks they accumulate.

Frequently Asked Questions (FAQ) about the Role of Citations in AEO

Do citations replace backlinks?

No. Citations in AEO differ from backlinks in SEO. They serve different functions within different systems, and both remain valuable.

Backlinks tell traditional search engines that other sites endorse your content, which influences your rank position in a list of results. Oppositely, citations tell AI answer engines that your content is the direct source behind a specific claim in a generated answer. But, you see, you need both because your audience is split across both discovery channels.

That said, here’s how they work together:

• Backlinks build domain authority that still drives organic rankings in Google’s traditional results. A strong backlink profile also contributes to the domain-level trust signals that AI engines consider when evaluating source quality.
• Citations earn you inclusion in AI-generated answers, where a growing share of buyers now start their research. They’re driven by content clarity, factual specificity, and structural readability, which are factors that backlinks alone can’t guarantee.

Citations in AEO are additive, not a replacement. Teams that abandon link-building in favor of citation-only strategies lose traditional search visibility. Teams that ignore citations while doubling down on backlinks become invisible in AI answers. The right approach is to run both in parallel.

Pro Tip: Use Marketing Hub with HubSpot AEO simultaneously to track performance across both channels — organic search traffic from traditional rankings alongside LLM referral traffic from AI citations. That dual view prevents you from over-indexing on either signal.

How long does it take to earn AI citations?

There’s no fixed timeline, but most teams can expect to see initial citation appearances within 4 to 8 weeks of publishing optimized content, with significant variation depending on three factors:

• Topical competition. Niche, specific queries with fewer competing sources get cited faster than high-volume, heavily covered topics. A detailed guide on AEO audit workflows will earn citations sooner than a generic “what is SEO” explainer.
• Content structure. Pages that use clear heading hierarchies, definition-style lead sentences, FAQ schema, and structured data are easier to discover.
• Domain trust baseline. Sites with existing authority and a track record of accurate, well-cited content get evaluated faster by AI engines. But the citation-quality data show that mid-tier sources account for nearly half of all citations, so a smaller domain with exceptional content specificity can outperform a larger one.

Can AI cite content behind a paywall?

In most cases, no. AI answer engines need to access and process your content to cite it, but hard paywalls block that access for both web crawlers and AI retrieval systems.

Here’s how different content access models interact with AI citation:

• Fully paywalled content (no access without login/payment) is effectively invisible to AI engines. It won’t be crawled, indexed for AI retrieval, or cited in generated answers.
• Metered paywalls (first few articles free, then gated) may allow AI engines to access and cite the free content, but anything behind the gate is excluded.
• Freemium models (full article visible, premium features gated) perform best for citation visibility because the core content is accessible while the conversion mechanism remains intact.
• Registration walls (free but requires email) vary; some AI crawlers can access this content, but many cannot.

Citations in AEO depend on your content being accessible to the systems generating answers. If your highest-value content is behind a hard paywall, it will not earn AI citations regardless of its quality.

Should I write for AI or humans first?

Write for humans first. Always.

The content attributes that AI engines reward are the same ones that make content genuinely useful to humans.

Every one of those qualities also makes content better for the person reading it:

• Clarity means a human can understand your point without re-reading the paragraph.
• Authority means you’re backing claims with data, experience, and specificity that a reader trusts.
• Structure means scannable headings, logical flow, and direct answers that respect a reader’s time.
• Freshness means current information that actually helps someone make a decision today.

The teams that try to “write for AI” are wasting their time by stuffing structured data, keyword-loading headers, and formatting content in ways that read awkwardly to humans, and end up producing pages that underperform with both audiences. AI engines are increasingly sophisticated at identifying content that prioritizes manipulation over genuine usefulness.

Write naturally for your human reader, then optimize the structure (headings, schema, lead sentences, factual density) for machine readability.

Pro Tip: Want a reliable gut-check test? Read your content aloud. If it sounds like a human expert explaining something to a colleague, it’s structured well for both audiences. If it sounds like a keyword list wearing a paragraph costume, AI engines will skip it just as quickly as human readers will.

How do I know if an answer engine cited my brand?

Tracking AI citations requires dedicated monitoring because they don’t appear in traditional SEO tools like Google Search Console or standard rank trackers. Here’s a full breakdown of what to track and how:

• HubSpot’s AEO Grader lets you input your target queries and see whether your content appears in AI-generated answers. This is the fastest way to benchmark your current citation visibility and identify gaps.
• Manual spot-checking across ChatGPT, Perplexity, Gemini, and Google AI Overviews for your priority queries. Run your top 10-15 target questions through each engine monthly and document which sources are cited.
• Brand mention monitoring across AI answers. Even when a citation doesn’t include a clickable link, AI engines may reference your brand by name. Tracking named mentions gives you a fuller picture of AI visibility than link-based citation tracking alone.

Citations in AEO make this tracking essential, not optional. Build citation tracking into your monthly reporting cadence alongside organic keyword rankings and traffic metrics.

Citations Are Just the Beginning of AEO Success

Citations are the most direct proof that your content is structured, authoritative, and current enough to be selected as an AI engine’s source of truth. But citations alone don’t capture the full picture.

Citations sit within a broader ecosystem of AI visibility metrics, which are:

• Brand mentions
• LLM referral traffic
• Answer inclusion rates
• Conversion from AI-driven visits

Together, they determine whether your content strategy is built for how buyers actually find answers today.

The good news? You don’t have to build this from scratch. HubSpot’s AEO Grader enables measurement of AI citation visibility, Content Hub gives you the structural foundation to publish citation-ready content at scale, and Marketing Hub connects AI referral traffic to the actual pipeline so you can prove ROI, not just report impressions. The infrastructure exists. The shift is happening. The only question is whether your content strategy moves with it.

Ready to see how your content performs in AI search? Get started with HubSpot’s AEO Grader today.

A few weeks ago, I wrote about our vision for the agent era: agents should be able to run on HubSpot, and to run HubSpot. I want to go a level deeper on what “run HubSpot” actually means, and our latest step in bringing this vision to life.

Businesses aren’t just sending employees into HubSpot to do work. They’re sending agents. And those agents need to be able to act as effectively as possible on your behalf, wherever they’re operating.

That last part is important. An agent isn’t always running in one place, on one infrastructure.

With AI Connectors, HubSpot context and actions are already available in Claude, ChatGPT, and other environments where teams work. Now, we’re adding another agent infrastructure: Command Line Interface (CLI).

Introducing the HubSpot Agent CLI

The HubSpot Agent CLI brings HubSpot’s data and intelligence into the environments where GTM and ops teams are composing their own workflows – Codex, Claude Cowork, and Claude Code – and allows agents to automate repetitive, bulk, and scheduled work.

The simplest way to think about it: take the questions you’ve been asking or the tasks you’ve been completing repeatedly in chat, and automate them. Build automations in Codex or schedule them in Cowork, and the work happens on its own before you even get to your desk.

It’s built on the same foundation as our public APIs and MCP server that already power our AI Connectors — and it’s designed to complement them, not replace them. AI Connectors are great for work where a human is in the loop, talking to an agent: questions, insights, conversations, campaign analytics. The Agent CLI can help agents accomplish those tasks, too, but it’s particularly useful for the repetitive, bulk, and scheduled work that needs to run without a human in the loop.

How the Agent CLI strengthens agents working on HubSpot

The HubSpot Agent CLI will help GTM and ops teams automate and schedule routine tasks, reports, and actions so they get more time back to do the work that matters. No more asking for the same thing multiple times. For example:

• A marketer could ask for a report every Monday at 8 a.m. that delivers high-fit contacts with no associated deal, no recent sales activity, or missing key enrichment fields, then send RevOps a prioritized cleanup list with suggested next actions.
• Sales and RevOps could have a daily scan of the pipeline for deals closing this week that have had no activity in the past five days, and ask for a summary.
• Customer Success could get an automated account review that summarizes open deals, recent support activity, and last NPS score for each account in the book of business.
• Support could set up an automation for whenever a new ticket comes in from a top tier account, the agent pulls the last five tickets from the contact, summarizes each resolution, and flags recurring issue patterns.

The work happens in the background, ready when you need it.

Why agent infrastructure optionality matters

We’re building a platform where agent infrastructure is a choice your agents make based on what’s right for your workflow.

An agent constrained to one infrastructure is less effective than it could be. Just as customers should have the freedom to choose the best tools for their business, agents should have the same. Optionality enables agents to choose the best infrastructure for the task at hand so they can operate most efficiently, whether they’re running a scheduled automation, processing a bulk operation, or acting on a real-time signal.

The direction is clear: wherever agents are working, and whatever infrastructure they’re running on, HubSpot supports it. That’s what building for the agent era looks like.

The HubSpot Agent CLI is available in private beta now, and anyone interested can sign up here.

Walk down a suburban street, and you might stumble across a following sign. It’s probably messy with poor formatting and inconsistent font size. Here’s one that I saw in Houston.

Source

One line stuck out to me. It reads, “Window cleaning in progress.” I love this board because it showcases something that most marketers forget. K&C Window Cleaning doesn’t try to persuade you with flashing slogans or in-your-face claims. They’re far more subtle, and that subtly makes them more effective.

While ads are about influence, no one wants to be sold to. Instead, this sign uses psychology to get people thinking, “Other people are using the service, so maybe it’s time to clean my windows.” The idea feels organic and nothing feels forced.

Table of Contents

We follow the actions of others

Back in 2008, the legendary researcher Robert Cialdini ran a notorious study. Set up over 80 days in a mid-priced hotel in the American Southwest, the three researchers ran tests in 190 rooms. Their goal was to encourage visitors to reuse their towels. Inside the room, they tested different signs with over 1,058 guests.

First, they tested a standard environmental message saying, “Help save the environment.” Guests said this message would be most likely to persuade them. But the researchers also tested a message that read “most guests reuse their towels.” The results were surprising.

The environmental plea encouraged 35% reuse, but the suggestion that the majority of guests reused their towels boosted reuse to 44%. But, then they added a third message: “Most guests in this room reuse their towels.”

This had an even greater impact. Now, almost 50% of guests reused old towels, up from 35% in the control. The takeaway is simple: we follow the actions of others.

So, if a neighbor pays for window cleaning, we’ll consider doing the same. But marketers forget one important element: Consumers don’t like to feel forced.

We don’t like to feel forced

Messages like “we’re the most popular” and “we’re number one” work, but they’re not perfect. Nicolas Guéguen in 2000 showed that people are more likely to act if they feel autonomous, not forced.

The study attempted to persuade French commuters to spare some coins for a bus ticket. The researchers tried two messages, which yielded surprisingly different results:

1. “Sorry, would you have some coins for me to take the bus, please?“ 10% agree
2. “Sorry, would you have some coins for me to take the bus please? But, you are free to accept or refuse.” 47.5% agree

This method, coined the “but you are free to refuse” technique, has been proven in multiple different domains, both online and offline. A 2013 meta-analysis found that the effect worked across 42 different domains.

That brings us back to K&C Window Cleaning’s sign. It takes all this advice to heart. It showcases the actions others take, but doesn’t force the reader into a corner.

• It doesn’t say, “We’re the most popular window cleaners in Houston.”
• It says, “Window cleaning in progress.”

And plenty of other companies do the same. They don’t say they’re popular; they prove it.

My favorite example comes from Sam Tatam’s wonderful book, Evolutionary Ideas. At his favorite cafe in Sydney, Australia, the owners don’t say, “We’re popular.” Instead, they show it by sticking the loyalty cards of their customers on the wall.

Don’t say it. Show it. It’ll make your message far more effective.

Make decisions feel natural

If you believe in your offering, you’ll want to brand it as the best, brightest, and most popular. Resist the urge. The best marketers let customers make their own decisions by showing value. The most persuasive thing you can do is make your customer feel like the idea was theirs all along, backed by peers who have also reaped benefits.

If you’re worried about what AI Overviews mean for SEO, let me remind you of the panic over featured snippets circa 2017. Remember how that turned out? At first, bloggers and SEOs bristled over these quick-glance summaries at the top of the Google SERPs, fearing they’d steal all our traffic. Eventually, however, we adapted and started optimizing content to get mentioned in them. I believe the same will be true of AI Overviews. I mean, it’s already happening: The internet is now filled with the latest advice on how to get cited in AI Overviews (including this article).

I wrote this guide for SEO and marketing leaders seeking practical frameworks to overcome declining clicks and optimize content for Google AI Overviews. Find out what triggers an AI Overview, how AIOs will change SEO playbooks, and where AI Overview optimization fits inside an existing SEO program. I’ll include lots of research and examples, too.

Source

A 2018 Reddit post about how featured snippets could “kill the regular internet”

A 2023 Facebook post from a blogger praising featured snippets and sharing how to capture them

Table of Contents

What are AI overviews?

AI Overviews are AI-generated summaries that appear at the top of the Google search results page, giving a direct answer to your query, synthesized from multiple websites. There may be a few images, links, and a “show more” button you can click to get more details. You can also click the cited sources to read those webpages.

Here’s an example of an AI Overview that the HubSpot Blog shows up in. Notice that HubSpot is featured four times above the fold: as two in-line citations within the AI Overview’s summary, as a clickable snippet to the right of the summary, and as the first blue link in the traditional search results below. Winning the AIO and first position means a brand can multiply the surface area it occupies in Google’s new SERP design.

AI Mode, on the other hand, is the full AI chatbot experience of Google Search. You can access it by clicking the “AI Mode” tab at the top of Google. Unlike the one-shot nature of AI Overview, AI Mode allows you to continue the conversation in multiple turns, similar to ChatGPT or Claude. The rest of this article will focus specifically on AI Overviews.

Importantly, the links that show up as sources in an AI Overview do not always overlap with the top 10 results on the SERP. That means your webpage can rank number one in Google and still get overlooked for an AI Overview. In fact, a Semrush study of 200,000 AI Overviews found that the number one search result appeared in the AIO only 34% of the time on mobile and 46% of the time on desktop.

Because AI overviews seek to instantly and directly answer the query, searchers often do not need to scroll or click further to get their answer. Their query is immediately satisfied. As you can imagine, this can have a devastating effect on click-through rates (CTR) and website traffic.

What AI Overviews Mean for SEO

AI Overviews correlate with lower click-through rates and higher zero-click searches, meaning you can expect to see less organic traffic from Google. On queries where AI Overviews appear, average outbound organic clicks dropped 38% and zero-click searches rose from 54% to 72%, according to a working paper published in April 2026 by researchers from the Indian School of Business and Carnegie Mellon University.

And AI Overviews are increasingly the norm: As of February 2026, they triggered on nearly half (48%) of tracked queries, according to BrightEdge. And when broken down by industry, that rate could be even higher: AI Overviews triggered on 84% of 1,000 B2B queries analyzed by AEO agency Fan Out.

From our own data, HubSpot customer organic traffic was down 27% year-over-year globally as of February 2026 (though I don’t know how much of that, if any, was due to Google AI Overviews). But what I’m saying is this: If you’re seeing consistent declines in traffic, you’re definitely not alone.

So, what do AI Overviews mean for SEO?

• Success metrics have to shift. With the rise of zero-click searches, clicks are a poor measure of success. Now, it’s about influencing buyers even when they never click through to your site. Instead of obsessing over keyword optimization, positions, and traffic, focus on brand visibility score, mentions, and citations.
• Traditional SEO still matters. AI Overviews are heavily influenced by the SEO fundamentals you already know: technical SEO, quality content, and topical authority. After all, Google has explained that “AI Overviews use a customized Gemini model, which works in tandem with our existing Search systems.” Ranking well in Google’s SERP will help you win AI Overviews (though it’s not guaranteed).
• Adding AEO is essential. Answer engine optimization (AEO) is separate from but complementary to SEO. Some AEO tactics have no SEO equivalent. For example, SEO has historically optimized for ranking on specific keywords, while AEO emphasizes topical breadth across the conversational, fan-out queries that answer engines generate behind the scenes. And while SEO is primarily focused on driving traffic to your site, AEO leans heavily on off-site presence — building entity signals through authentic mentions on platforms like Reddit, YouTube, and industry publications — so your brand surfaces in AI answers even when no one clicks through.

What Triggers an AI Overview

Not every query will trigger an AI Overview in Google Search. Here’s what we know triggers an AI Overview, based on large-scale studies from Ahrefs and Semrush.

1. Informational Intent Keywords

The majority of AI Overviews come from informational intent keywords (from top-of-funnel users who just want to learn), but over the past year, AIOs have been moving down the funnel, according to the latest Semrush data. Semrush’s analysis of 10M+ keywords found that the share of informational queries triggering AI Overviews dropped from 91.3% in January 2025 to 57.1% by October 2025. Over the course of 13 months:

• Commercial queries grew from 8.15% to 18.57%. These indicate mid- to bottom-of-funnel users evaluating a potential purchase.
• Transactional queries grew from 1.98% to 13.94%. These indicate bottom-of-funnel users looking to make a purchase immediately.
• Navigational queries skyrocketed from 0.84% to 10.33%. These are typically from users who are Googling a name to reach a website.

Informational pages are still the most likely AIO targets, but the gap is closing fast. Don’t assume your comparison, pricing, or branded pages are safe from AIO disruption because, increasingly, they aren’t.

2. Questions, Especially Those Starting With “What,” “How,” and “Is”

Ahrefs found that 57.9% of all question queries triggered an AIO. Further, Semrush found that among the question keywords that triggered AI Overviews in its sample, those starting with “what,” “how,” and “is” appeared most frequently.

3. Queries about Science and People & Society

Science and People & Society are consistently among the industries most likely to trigger an AI Overview, per both Ahrefs and Semrush. Ahrefs found that 43.6% of Science queries and 43.0% of Health queries triggered an AIO — more than double the 21% baseline across all keywords. Semrush’s analysis of November 2025 data placed Science, Computers & Electronics, and People & Society among its top-cited industries.

4. Long Queries (7+ words)

The longer a query, the more likely it is that an AI Overview will appear. Forty-six percent of queries that are seven or more words trigger an AI Overview, according to Ahrefs data. The research also found that the chances of an AIO appearing increase incrementally as query length increases, starting at 9.5% for one word and maxing out at 46.4% at 7+ words.

5. Non-Branded Queries

Ahrefs found that non-branded queries are 1.9x more likely to trigger an AIO than branded queries (24.9% versus 13.1%). Here’s an example: When I enter the non-branded query of “calorie tracking app,” I get this AIO:

But when I enter the branded query of “MacroFactor,” I don’t get an AIO at all. Instead, I get MacroFactor’s website.

This makes sense when you think about intent (which we talked about earlier): Someone typing in “MacroFactor” probably has navigational intent — they’re trying to get to that specific brand’s website. But someone typing in “calorie tracking app” likely has informational or maybe even commercial intent — they’re trying to get more information about an app and/or they’re considering buying.

Overall, if an AI Overview appears for a query, users are far less likely to click any links. That means the goal now is to get cited in the AI Overview to win the visibility and traffic you can. The next section will show you how to do just that.

How to Get Cited in Google AI Overviews

Let’s start with the bare minimum, without which you won’t be able to show up in AI Overviews or Google Search at all:

• Your site can’t be blocking GoogleBot, Google’s crawler.
• Your content shouldn’t be in violation of any of Google’s policies.
• The page should load (return an HTTP 200 success code as opposed to, for example, a 404 error).

Google insists that, beyond those listed above, there are no further technical requirements to be eligible for an AI Overview. However, what the search giant does not explicitly share is how to optimize for AI Overviews (i.e., increase your chances of getting chosen for an AIO).

For that, we’ll need to turn to industry experiments and best practices for AI search content. Most of the evidence below is based on actual analysis of thousands of AI citation data points reported in HubSpot’s State of AEO 2026. I’ll also cite more of the research from the report I mentioned earlier from AEO agency Fan Out.

1. Focus on publishing blog content.

Across eight content types, blog posts/informative articles are the most-cited for AI Overviews, with a 42% citation rate, according to HubSpot’s State of AEO 2026. The least cited content types in AI Overviews are news (non-evergreen articles) at 5% and PR at 6%.

This makes sense based on the other content types analyzed in the report and what we already know about what triggers AI Overviews: 57.1% of queries that trigger AIOs are informational, and compared to other content types — like comparisons, which lean toward commercial intent and are favored by ChatGPT — blog posts are highly informational. So, if you’re specifically trying to win AIOs, blog posts should be your focus.

2. Optimize your titles.

Across eight title patterns, “What is [X]” is the top performer for AI Overviews, while “Best [X]” lists and How-tos work well too. Including the year in H1s and meta titles also correlates with higher citation rates in AIOs.

Ideal title for AI Overview: “What is the best website builder for beginners in 2026?”

Not ideal title for AI Overview: “The complete guide to website builders for beginners”

3. Add FAQ sections to your pages.

Pages with FAQ sections are more likely to be cited in AI Overviews, according to the State of AEO 2026. Adding schema markup (a type of structured data you add as code snippets) to those FAQ sections correlates with higher citations in Google AI Mode, Gemini, and Perplexity. Now, having said that, let me be clear: AI Overviews do not require schema markup.

Below is what a good FAQ section looks like, courtesy of HubSpot’s Content Hub pricing guide. Notice the descriptive H2 heading (“Frequently Asked Questions About Content Hub Pricing”) and the questions formatted as H3s — the State of AEO report found that this combination could provide a citation boost.

4. Build EEAT signals in each page.

AI Overview is among the most responsive to EEAT signals compared to any of the other answer engines the State of AEO report analyzed. It makes sense, given that the EEAT framework came from Google.

Specifically, the State of AEO found the following page elements are tied to an increase in AI citations (so adding these to your content might help you get cited in more AIOs):

• Outbound links (especially true for AIOs and Gemini)
• Statistics and data (especially true for AIOs and ChatGPT)
• Author bio on page (slightly higher citation impact than the author name alone)
• A visible “Last updated” date (a stronger citation predictor than the original publish date)

Let’s dissect an excellent example of building EEAT signals into a page from NerdWallet, a company that State of AEO’s analysis identified as one of the most cited B2C brands in its dataset. NerdWallet wins the AIO for the highly competitive query “how to track expenses,” showing up not once, but twice.

Now, let’s click through to the blog post and see how NerdWallet nails EEAT signals on-page. First, there’s an author byline which, upon hover, triggers a pop-up with the author bio (the author bio is also available at the bottom of the article). Within that bio, there’s a wealth of credibility signals: the author’s years of experience, areas of expertise, and even the publications she’s been published in. On top of that, the page clearly displays when the post was last updated.

Source

5. Apply traditional SEO principles too.

And last but certainly not least is our old friend, SEO. No, it hasn’t gone anywhere. In fact, a strong SEO foundation is essential to building a successful AEO program — and SEO expert predictions point to that same overlap between traditional ranking and AI visibility.

I spoke with Elie Berreby, head of SEO and AI search at Adorama, which was named “Top Overachiever” in consumer electronics AI brand visibility by Similarweb in 2026. This means Adorama “ranks significantly higher in AI visibility than in branded search demand,” according to Similarweb’s report. Berreby reiterated the interdependence of traditional search and AI Overviews. “If you are not optimizing for Googlebot, obviously, you are ignoring the entire Google ecosystem, and you’re going to suffer in Gemini, in AI Mode, and in the AI Overviews.”

According to the State of AEO report:

• From the dataset, HubSpot found that pages that rank high in Google are more likely to be cited in an AI Overview than other answer engines.
• Pages that rank for more Google search terms and rank highly in SERPs are more likely to be cited by answer engines in general.
• With an ideal keyword range of 50+, AIO was tied with Perplexity for the highest preferred number of keywords of any of the AI answer engines in the dataset (including Gemini, AI Mode, Copilot, ChatGPT, and SearchGPT).

Takeaway: If you want a page to win an AI Overview, cover more subtopics and answer more subquestions in that one piece than you would have before.

Pro tip: To write a comprehensive piece that’s AIO-worthy, AEO strategist Kaleigh Moore recommends owning the topic by thinking holistically about the questions your buyers might ask. “How do we think about topical ownership for our specific customer trying to solve this very specific problem?” Moore says. “What are the types of questions they’re asking at this stage of the buyer’s journey? And how do we create content that proactively answers those questions?”

6. Create content for and get mentioned on Reddit and YouTube.

For off-site strategy, YouTube is the first place marketers interested in appearing in more AI Overviews should invest. Sixty-one percent of YouTube citations in Fan Out’s analysis came from AIOs. And given that Google owns YouTube, I’m not at all surprised that AIO prefers this platform most. To put that favoritism in perspective, of the 1,000 queries in Fan Out’s dataset, ChatGPT Plus cited no YouTube links at all.

Here’s some of what Fan Out’s report reveals and recommends for optimizing YouTube for AEO:

• Comparisons and tutorials dominate, so create content in these categories. Specifically, Fan Out names problem-solution, comparison, and best-of as the top three YouTube video types.
• 13.7% of the YouTube citations had timestamps, so Fan Out recommends adding chapter markers and timestamps to your videos.
• Optimize your video descriptions to help LLMs understand what your video is about.

In my own experience, AIOs almost always cite a YouTube video for “how to” queries, whether that’s “how to bake banana bread:”

Or “how to use a crm for sales.”

Therefore, I recommend that if you want more AIO citations, create YouTube content targeting “how to” queries, or even partner with YouTube influencers to create content on behalf of your brand.

Reddit is the second-best place to optimize for AIOs. In Fan Out’s analysis of over 33,000 AI citations across Perplexity, Google AI Overviews, and ChatGPT Plus, Reddit surfaced as the #1-cited off-site platform overall. However, Reddit was by far most favored by AIO, which made up a 51% share of the Reddit citations (substantial, but still less than YouTube).

Fan Out data reveals this about Reddit:

• Posts are far more likely to be cited by AI engines than comments.
• The top three content types are best-of, alternative-seeking, and review-opinion.
• Engagement on a Reddit post doesn’t matter nearly as much as content structure and topical relevance.

Here’s the power of Reddit on AIOs in real life: I Googled “best crms for small business” and HubSpot was recommended first, thanks to a Reddit post that HubSpot didn’t even create or comment on.

Measuring and Diagnosing AI Overview Impact

Queries that led to AI Overviews are included in the overall performance report in Google Search Console, but they’re aggregated with the rest of your SEO keywords, so you cannot filter to see only the metrics related to AIOs. To do that, you’ll have to use a third-party tool. Here are two that I recommend.

Ahrefs Brand Radar

Best for: answering the question, “What queries are triggering AI Overviews for this specific page or domain?”

Enter any URL and Ahrefs Brand Radar shows you the queries where it was cited in an AI Overview, including an estimated average monthly search volume for the query, the content of the AI Overview, and a list of the cited domains included in that AIO.

If you’re already using Ahrefs for SEO research and tracking, then it makes sense to use its Brand Radar too for the AEO piece.

Pricing: Brand Radar is included in paid Ahrefs subscriptions, which start at $129/mo, but there are limitations on the number of prompts you can track. Alternatively, Brand Radar is available as a standalone tool starting at $199/mo for one AI platform (such as AI Overviews/AI Mode) but doesn’t include custom prompt tracking. To get custom prompt tracking, you can purchase an add-on package.

Otterly

Best for: answering the question, “Is this domain getting cited in AI Overviews for the tracked prompts I care about?”

Otterly tracks prompts across six answer engines, including Google AI Overviews, surfacing both citation data and brand sentiment. I’ve tested Otterly quite a few times for other articles, including one on the best AEO rank trackers, and I really appreciate its ease of use and the fact that it’s a dedicated AEO tool (as opposed to doing both AEO and SEO). For brands focused on AI visibility, Otterly will be more straightforward and affordable than Ahrefs Brand Radar.

Pricing: Otterly pricing starts at $29/mo for 15 tracked prompts across ChatGPT, Google AI Overviews, Perplexity, and Microsoft Copilot.

Frequently Asked Questions About AI Overviews and SEO

Can you fully opt out of appearing in AI Overviews?

No, you cannot opt out of having your website appear in AI Overviews specifically, but there is a workaround: Inserting the “nosnippet” robots meta tag prevents pages from being included in snippets not only for AI Overviews/AI Mode but also for Google’s traditional search (e.g., it wouldn’t show up for a featured snippet).

• Page-specific. Paste this code between and of the HTML on every individual page that you don’t want appearing in AI Overviews.

  Code:

• Sitewide: If your site uses a template, you can paste this code in the template header if you don’t want any of the pages that use that template to appear in AI Overviews:

  Code:

Pro tip: Do not mix up “nosnippet” with “noindex.” Adding the “noindex” robots meta tag would prevent the affected pages from showing up in all of Google Search (not just AI Overviews), so it is extreme, ill-advised, and would hurt the affected pages’ SEO.

Now, on the flip side, as a searcher, you can’t really opt out of seeing AI Overviews when you perform a search in Google. Here’s what Google says in its help center: “AI Overviews are a core Google Search feature, like knowledge panels. Features cannot be turned off.” However, there are a couple of workarounds:

1. Add “-ai” at the end of your query. This usually prevents AI Overviews from appearing in your Google search results, but it’s not guaranteed.
2. Click the “Web” tab filter at the top of the Google results page. This will show you only the web results, minus the AI Overviews.

How do you track clicks from AI Overviews specifically?

Right now, it isn’t possible to reliably isolate clicks from AI Overviews in Google Search Console. The best you can do at the moment is track AIO visibility, and to do that, you’ll need a third-party tool. Ahrefs Brand Radar supports AI Overview coverage and can help you find queries where AI Overviews mention your brand, cite your website, or both. From there, you can compare those visibility signals with GSC and analytics data to estimate the impact AIOs are having on your organic traffic.

Should you create pages specifically for AI Overviews?

No. Google says there are no additional requirements or special optimizations needed to appear in AI Overviews, and both Google and Bing representatives have cautioned against creating separate Markdown, JSON, or bot-facing versions of pages just for LLMs. Instead, focus on improving your existing content using the practices covered above: clear definition-style answers, FAQ sections, EEAT signals, and a visible “Last updated” date. That means optimizing for both traditional SEO and AI search visibility simultaneously, without creating separate “AI-only” pages.

How do you brief executives on AI Overview impact?

When briefing executives on the impact of AI Overviews on your website and business, reframe the challenge as an opportunity to win mindshare and shift the success metrics from clicks to visibility.

“Our site is losing traffic” is a bad look for any marketing org. But ultimately, leadership cares about the bottom line: Is this earning the business money or not? Clicks used to be a proxy for that, but they’re not anymore. It doesn’t matter if your site is losing clicks if the business is actually gaining customers.

Studies show that LLM traffic converts at a higher rate than traditional Google-referred traffic. Ahrefs found that its AI search traffic converts 23x better than traffic from traditional organic search. Semrush concluded that LLM visitors converted 4.4x better than traditional organic search visitors, after analyzing over 500 digital marketing and SEO topics.

Suggest a shift from traffic to share of voice; that’s the metric that shows how your brand stacks up against competitors in AI visibility. Because sure, your site might be losing traffic, but there’s a good chance that across the board in your category, competitors are losing traffic too. What matters, then, is that when a potential customer asks about your category, whether they consult Google AI Overviews, ChatGPT, or Perplexity — your brand is the one they see. Communicate to leadership that by winning more AI Overviews (even if your brand isn’t linked or your site isn’t clicked), you’re actually winning mindshare.

Beyond AI Overviews: Tracking Your Visibility Across All Answer Engines

By now, I hope you see that AI Overviews, like the featured snippets that preceded them, represent a push-and-pull that has always existed between content creators and Google: The search engine constantly changes its algorithm and SERP design, and what else can we do besides adapt? AI Overviews could be an exciting opportunity to be the recommended answer to a potential customer’s question — even if they never click on your website.

I do want to caution against taking a myopic view of AEO and focusing only on AI Overviews, however. They are just one surface in the AI search landscape. You can’t forget about the opportunities for your brand to show up in answers from other LLMs, such as ChatGPT and Perplexity.

HubSpot AEO is a specialized tool that tracks citations in ChatGPT, Perplexity, and Gemini and provides recommendations to increase your visibility in each. Pricing starts at $50/month with no other HubSpot subscription required.

For teams running marketing in HubSpot already, AEO is included in Marketing Hub Pro and Enterprise. The integrated version uses your CRM data to suggest which prompts to track from day one, so tracking is grounded in your actual business context rather than generic category guesses. And because the recommendations connect to HubSpot’s content tools, you can take action on AEO insights within the same tool.

The takeaway: AI Overviews matter, but they’re only one slice of the bigger AEO opportunity. Once you broaden the lens to every major answer engine, you can stop guessing where your visibility is leaking and start closing those gaps systematically.