How to Structure Content So AI Search Cites Your

Every few weeks, a new video promises the definitive trick for ranking in AI search. A few weeks after that, the same creator is back with a different trick, sheepish about the last one. If you've been in digital marketing for more than a month, you've watched this cycle spin a few times already.

Brian Dean, who founded the trend-tracking platform Exploding Topics, has a pointed take on this pattern: he says it's largely noise, and that Exploding Topics ranks for thousands of prompts across ChatGPT, Gemini, and Claude without using any of it. Speaking in a recent video for Semrush, Dean lays out a four-part framework for AI search visibility — and the notable thing isn't how exotic it is. It's how much it rhymes with what good web publishing has always looked like.

That's either reassuring or suspicious, depending on your priors. Let's walk through it carefully.

The reframe: AI as a second channel, not a replacement

Dean opens with a concession that a lot of SEO commentators are still reluctant to make clearly: Google traffic is down. "Some searches that used to send clicks are now being answered directly in tools like Google AI Overviews or ChatGPT," he says. For informational queries — the "what is X" and "how does Y work" type — AI tools have become effective answer machines that don't always need to send users anywhere.

His response to this isn't to panic or declare SEO dead. It's to reframe the math. If you're losing some informational traffic to AI Overviews, you can offset that loss by being the source AI tools cite when users ask relevant questions. "We now have a second channel to work with, and that's AI," Dean says. "A lot of our new leads and customers tell us that they found us using AI tools like ChatGPT and Claude."

This framing — AI as additive traffic channel rather than pure cannibal — is becoming a common position among practitioners who've moved past the initial alarm. Whether that optimism holds as AI tools get more capable at retaining users is an open question, but for now it appears to be holding for sites like Exploding Topics.

The generative engine optimization conversation has been circling this premise for months: that the right response to AI search isn't defensive but adaptive.

Step one: Stop chasing informational keywords

Dean's first tactical move is a reallocation of content effort. He's not saying abandon informational content — he's saying stop prioritizing it. The reasoning is straightforward: AI tools are genuinely excellent at answering informational queries. They synthesize, summarize, and answer "what is a CRM" better than most blog posts do, without needing to send the user anywhere. Competing in that space is increasingly a losing proposition.

Where AI tools are weaker, and where they still tend to cite external sources, is in high-stakes decision contexts. Queries like "best trend forecasting tools," "Exploding Topics alternatives," or "market research software" involve comparison, judgment, and implied purchasing intent. Users in this mode are more likely to click through to sources. AI tools, serving users who clearly want to evaluate options, are more likely to name them.

There's a coherent logic here that's worth taking seriously. Informational content still serves a purpose — Dean acknowledges it helps build topical authority and gives AI systems more context about what your brand covers. But as a primary traffic strategy, it's increasingly a bet on diminishing returns.

The harder question this raises: if everyone shifts to buyer-intent content, does that space get as saturated as informational content already is? Possibly. Dean's framework works in part because it's not yet the default playbook.

Step two: Chunk your content like each section is its own article

This is the most concrete and transferable piece of Dean's framework. His term for it is "chunking" — structuring content so that each section is independently intelligible and directly answers a specific question.

The mechanics are simple: use a subheading specific enough to function as an article title on its own ("How to Manage Seasonal Inventory: A Step-by-Step Process" rather than "Step 2"). Follow it immediately with one or two direct sentences that answer the implied question before any elaboration. Then add supporting proof — data, examples, pricing comparisons, expert input.

The underlying logic is about how AI systems extract and cite information. A language model scanning a page for a citable answer is doing something functionally similar to a human skimming for the relevant section. If the structure is vague, the model either skips it or miscontextualizes it. If each section is self-contained and clearly labeled, it becomes a clean, extractable unit.

"This structure makes your page easy to scan for readers and easy for search and AI systems to pull from," Dean says. What's notable is how well this aligns with accessibility best practices, plain-language writing guidelines, and what Google's own quality guidelines have recommended for years. Google's guidance, as Dean notes, explicitly says traditional SEO still matters for AI search because AI features rely on existing quality and ranking systems. There's no secret decoder ring here — good structure is good structure.

Ahrefs has made similar observations about content formats that win AI citations, and the pattern is consistent: discrete, well-labeled, answer-first writing outperforms dense narrative in AI retrieval contexts.

Step three: Make your brand legible as an entity

Dean's third step is what he calls "entity authority" — the practice of making it consistently clear, across your most important pages, what your brand is and what it does.

For Exploding Topics, the core identity statement is simple: "We help people find trends early." Dean argues that statement needs to appear coherently on the homepage, the about page, and the product pages. Not necessarily verbatim, but consistently enough that an AI system training on or indexing those pages arrives at a stable understanding of what the brand is.

The reasoning is essentially about AI confidence. "AI tools need to understand your brand before they can confidently cite you in answers," Dean says. "The more they see the same thing cited about your brand on your site, the more confident AI tools will be about what you do."

This is plausible — AI systems do construct and reinforce entity associations through repeated signal. It's also the kind of advice that sounds almost obvious once stated, which is maybe the point. A lot of websites are genuinely inconsistent about how they describe themselves, using different positioning language across pages without thinking much about it. Tightening that up costs nothing and probably helps in multiple contexts beyond AI search.

Step four: Get other sites to talk about you

The final piece is the least controllable and the most important: third-party mentions. Dean argues that AI tools weight what other sources say about you heavily — Reddit threads, industry blogs, newsletters, journalism. These are the signals that corroborate your own on-site claims.

His strategy for earning them at Exploding Topics is to publish original data. Instead of pitching journalists on covering the product, the pitch is the data: "Here are the fastest-growing AI startups this quarter." Journalists and bloggers cite the data, and in doing so, cite the brand. Dean says this generates dozens of such mentions monthly.

There's nothing new about data-driven link-building as a PR tactic — it predates the AI search era considerably. What Dean is pointing to is that this strategy has gained a second layer of value: those mentions don't just help Google rankings, they feed the training data and real-time indexing pipelines that AI tools use to determine who's worth citing.

This is where the framework gets interesting from a structural standpoint. If AI citations increasingly depend on third-party mentions, and those mentions increasingly require original data worth citing, then the brands with the most citeable insights have a durable advantage. Exploding Topics is in an unusual position here — their core product is data about trends, so producing citeable trend data is essentially free marketing. Most brands don't have that natural alignment, and replicating the strategy requires either a significant research investment or a genuinely distinctive angle.

Dean acknowledges this implicitly. The strategy that's "worked best for us" at Exploding Topics leverages a structural advantage most sites won't share. That doesn't invalidate the approach — data-driven PR is genuinely effective — but it's worth noting that the playbook is easier for some brands than others.

The broader picture here is that Dean's framework is, at its core, a case for doing the unsexy fundamentals well: write for real intent, organize your content clearly, define your brand consistently, and earn external credibility through usefulness rather than outreach. That these principles apply to AI search as much as they apply to traditional SEO either means the fundamentals are genuinely durable — or that the AI search era hasn't yet diverged enough from the Google era to require a different kind of thinking.

Which of those is true matters enormously. We won't know for a while.

By Marcus Chen-Ramirez, Senior Technology Correspondent, Buzzrag