Every few years
there’s a moment in search where the teams that understood what was changing built something that compounded for the next decade, and the teams that waited until the change was undeniable spent that decade catching up.
Thank you for reading this post, don’t forget to subscribe!The early adopters of structured data in 2012 are still benefiting from the authority signals they built then. The teams that understood E-A-T before the Helpful Content Update weren’t scrambling to retrofit expertise signals into content they’d already published. The publishers that built topical authority through hub-and-spoke architecture before everyone else discovered it held rankings through algorithm updates that shook everyone else.
We’re in one of those moments right now. The window isn’t closed yet. But it’s closing faster than most content teams are moving.
What Has Actually Changed
AI Overviews are eating the top of the funnel. Google’s AI Overviews — the generated summaries that appear above organic results for informational queries — are reducing click-through rates on the queries they cover, even when the ranking holds. Google I/O 2026 confirmed what the traffic data had been suggesting: the searches that used to produce a click to your blog post are increasingly producing an AI-generated answer that keeps the user on Google. The content that gets cited in those overviews gets a different kind of visibility. The content that doesn’t gets less traffic than its rankings would have historically produced.
Agentic AI is starting to use content operationally. The next generation of AI tools — agents that browse, research, and complete multi-step tasks — use content differently than search users do. They’re looking for specific, structured, verifiable information that they can incorporate into a workflow. Content that’s built for agents — clearly structured, semantically precise, with schema markup that makes it machine-readable — is positioned differently in this environment than content built purely for human readers browsing search results.
Generative Engine Optimization has become a real discipline. GEO — the practice of structuring content so that AI systems cite it, reference it, and surface it in generated responses — isn’t a replacement for SEO. It’s an additional layer of optimization that the content operations building it now are establishing early authority in. The brands that appear consistently in AI-generated responses on their target topics are building a different kind of presence than the ones optimizing only for traditional search rankings.
These three shifts point in the same direction: the content that wins in the next phase of search is not the content that technically satisfies a query. It’s the content that genuinely deserves to be cited — that has original perspective, verifiable expertise, and structural clarity that makes it usable by both human readers and AI systems.
What to Build Right Now
Original data and proprietary perspective. The single most durable content investment you can make right now is content that contains information that doesn’t exist anywhere else. Primary research. Client data (anonymized). Observations from direct practice that can’t be replicated by an AI trained on published content. Information gain — the degree to which a piece adds to the existing conversation rather than summarizing it — is becoming the primary differentiator between content that gets cited in AI responses and content that doesn’t.
Entity authority for your brand and your people. Search engines now evaluate entities — brands, people, concepts — not just documents. Your brand’s entity status, your key contributors’ entity status, and the relationships between your brand and the topics you claim expertise in are signals that traditional SEO metrics don’t capture. Building structured data that clearly asserts who you are, what you do, and what topics you’re authoritative on is infrastructure work that most content teams haven’t done yet.
Schema markup as a first-class production step. Schema markup is no longer a technical SEO afterthought. It’s the language that makes your content legible to AI systems that are increasingly mediating between your content and the people who need it. FAQPage, Article, Person, Organization — the schema types that describe what your content is and what it contains are the same schema types that AI systems use to evaluate whether your content is worth citing. Implementing them consistently is table-stakes infrastructure for the next phase of search.
Topical authority at depth, not breadth. The content operations that survive AI Overviews are the ones that own a topic deeply enough that the AI has to cite them. Broad coverage of many topics produces a lot of content that AI can summarize without citing. Deep coverage of a specific domain — the definitive resource on a specific set of questions, with interconnected content that reinforces and cross-references — produces an authority signal that AI systems respect. Hub-and-spoke architecture isn’t new. Its importance in the current environment is.
A newsletter that isn’t dependent on search. The traffic risk from AI Overviews is real and it affects informational content disproportionately. The content operations with owned audience — email subscribers, direct relationships with readers — are insulated from search volatility in a way that search-dependent operations aren’t. If your content operation doesn’t have a meaningful email list, building one is the highest-priority distribution investment you can make right now. Not eventually. Now.
What Not to Do
Stop producing commodity content — posts that cover topics the same way they’ve been covered everywhere else, without original perspective or specific expertise. This content was marginal before AI Overviews. It’s invisible now. Every hour spent producing it is an hour not spent on the original, specific, authoritative content that actually compounds.
Stop treating schema as optional. It was optional in 2020. It’s infrastructure now. Every post that publishes without appropriate schema markup is a missed opportunity to signal to AI systems what the content is and why it’s credible.
Stop optimizing for the search experience that existed two years ago. The ranking factors, the click patterns, the user behaviors — they’re shifting. The content operation that’s still optimizing for 2022 search is building for a world that’s already changed.
The Window
Early-mover advantage in search has always been real and always been temporary. The teams that built topical authority, entity signals, and schema infrastructure before those things were standard practice built assets that have compounded for years. The teams that waited until those things were well-understood built adequate content operations, not exceptional ones.
The current moment is another early-mover window. The content teams that understand what GEO, agentic search, and entity authority require — and build for those requirements now — are establishing positions that will be significantly harder to challenge once everyone else has caught up.
The window isn’t closed. But it’s not as wide as it was six months ago.
FAQ: Content Strategy for the Next 18 Months
Original data and proprietary perspective — content that contains information that doesn’t exist anywhere else. Primary research, client data (anonymized), observations from direct practice that can’t be replicated by an AI trained on published content. Information gain is becoming the primary differentiator between content that gets cited in AI responses and content that doesn’t.
GEO is the practice of structuring content so that AI systems cite it, reference it, and surface it in generated responses. It’s not a replacement for SEO — it’s an additional layer of optimization. The brands appearing consistently in AI-generated responses on their target topics are building early authority. So yes, your content team does need to care about it.
AI Overviews reduce click-through rates on informational queries even when rankings hold. The content that gets cited in those overviews gets a different kind of visibility. The content that doesn’t gets less traffic than its rankings would have historically produced. The response is building content original enough to deserve citation — not trying to optimize around the overviews themselves.
Not instead — alongside. The decay signal in your existing archive is real and worth managing. But stopping new production to focus entirely on updates misses the original data and proprietary perspective that the current environment rewards. The answer is a measurement framework that identifies which existing content needs updating, freeing production capacity for the original content that compounds.
If your content operation is producing without a point of view, that’s the problem CCA solves. Start with the content audit.