How AI-driven search changes strategy, and what to adapt to keep visibility and traffic.
The fastest-moving story in search right now is also the most misread. Google's rollout of AI Overviews — the generated answer block that now appears above the blue links for roughly half of US queries — is being framed as a traffic apocalypse for publishers. The data tells a more nuanced story. Total Google traffic to most B2B sites is down somewhere between 8 and 25 percent year over year. Specific query types are down 50+ percent. But a small set of pages are up, sometimes dramatically. Understanding the difference is the work.
For the last two decades, Google was a query-to-link machine. You typed a question, Google returned ten links, you clicked one. The search engine's job ended at the click. Whatever ranking signals Google used — backlinks, on-page relevance, freshness, click data — fed into that one job.
AI Overviews changed Google's job. Now the search engine's job is to answer the question, not just route to it. The model behind AI Overviews reads multiple ranking pages, synthesizes an answer, and presents it inline with citations. For informational queries, this is often where the user's journey ends. They got their answer; they didn't need to click.
That sounds bad for publishers and it is, for one specific kind of content: generic, summary-style articles that exist to define a concept. "What is account-based marketing?" "How does retargeting work?" "What is a buyer persona?" These pages are exactly what AI Overviews can compress into two paragraphs without the user clicking anywhere. If your traffic was concentrated on this kind of query, you've already seen the drop.
Three categories of content are growing in the AI Overviews era. The first is original data. When someone searches "average B2B email open rate 2026," the AI Overview will cite a publisher who actually ran the analysis. The cited publisher gets the brand mention and frequently the click — because users want to verify the number, see the methodology, or pull the raw data. Original data is uncompressible. AI Overviews can summarize a number; they cannot replace the underlying study.
The second is expert opinion with a clear point of view. "Should I use HubSpot or Salesforce for a 50-person B2B SaaS?" is a comparison query, and AI Overviews try to handle it. But comparison answers benefit from a sharp human judgment — "if you're under 100 reps and not yet on enterprise contracts, the answer is HubSpot, and here's why we recommend this even though Salesforce's marketing pitch is stronger." That kind of opinionated answer is hard for a synthesis model to produce, and it's what users click through to read.
The third is utility — calculators, templates, frameworks, interactive tools. A model can describe a discount-cash-flow calculator. It cannot run one. "Calculate my marketing ROI" or "build a sales forecast" queries land on pages that do the thing, and those pages are growing.
One of the less-discussed shifts is how AI Overviews change link economics. In the link-based ranking era, what mattered was the inbound hyperlink — the explicit HTML <a href> from one site to another. AI Overviews are increasingly citing brands by name even without a clickable link, and the underlying ranking model is now measuring how often your brand is mentioned in trusted sources across the web. Forum discussions, podcast transcripts, video descriptions, expert roundups — they all feed the same signal: is this brand recognized as an authority on this topic?
This means PR, podcast appearances, and expert contributions to industry publications are now SEO work in a way they weren't before. A guest essay in an industry trade publication that doesn't even include a link back can still meaningfully improve how often your brand appears in AI Overview citations for your category.
If you're operating SEO in 2026 the way you operated it in 2022, you're losing ground. Three structural changes are worth making this quarter.
First, audit your existing content portfolio by query type. Pages that target generic definition queries should be deprioritized for new investment — they're shrinking targets. The traffic on those pages is going to AI Overviews and is not coming back. Reallocate effort to the three categories above.
Second, build at least one original-data piece per quarter. A 200-person survey, a benchmark across your client base, or a teardown of 50 examples in your category. These take a quarter to produce and earn citations for years.
Third, start treating brand mention as a tracked output. Tools like Brand Mentions, Talkwalker, and Ahrefs's brand monitoring can show you how often your category-defining terms get associated with your brand in trusted sources. That's the new equivalent of "backlinks" — the measurable surface of authority.
AI Overviews didn't end SEO. They ended the version of SEO that competed on producing roughly the same article slightly better than the existing top result. The new version of SEO competes on things that are genuinely uncompressible: data only you have, opinions only your experts hold, tools only your code runs, and brand mentions only your category presence earns. The teams that adapt early will compound an advantage. The teams that treat AI Overviews as a temporary disruption will spend three years losing ground they won't get back.
The fastest-moving story in search right now is also the most misread. Google's rollout of AI Overviews — the generated answer block that now appears above the blue links for roughly half of US queries — is being framed as a traffic apocalypse for publishers. The data tells a more nuanced story. Total Google traffic to most B2B sites is down somewhere between 8 and 25 percent year over year. Specific query types are down 50+ percent. But a small set of pages are up, sometimes dramatically. Understanding the difference is the work.
For the last two decades, Google was a query-to-link machine. You typed a question, Google returned ten links, you clicked one. The search engine's job ended at the click. Whatever ranking signals Google used — backlinks, on-page relevance, freshness, click data — fed into that one job.
AI Overviews changed Google's job. Now the search engine's job is to answer the question, not just route to it. The model behind AI Overviews reads multiple ranking pages, synthesizes an answer, and presents it inline with citations. For informational queries, this is often where the user's journey ends. They got their answer; they didn't need to click.
That sounds bad for publishers and it is, for one specific kind of content: generic, summary-style articles that exist to define a concept. "What is account-based marketing?" "How does retargeting work?" "What is a buyer persona?" These pages are exactly what AI Overviews can compress into two paragraphs without the user clicking anywhere. If your traffic was concentrated on this kind of query, you've already seen the drop.
Three categories of content are growing in the AI Overviews era. The first is original data. When someone searches "average B2B email open rate 2026," the AI Overview will cite a publisher who actually ran the analysis. The cited publisher gets the brand mention and frequently the click — because users want to verify the number, see the methodology, or pull the raw data. Original data is uncompressible. AI Overviews can summarize a number; they cannot replace the underlying study.
The second is expert opinion with a clear point of view. "Should I use HubSpot or Salesforce for a 50-person B2B SaaS?" is a comparison query, and AI Overviews try to handle it. But comparison answers benefit from a sharp human judgment — "if you're under 100 reps and not yet on enterprise contracts, the answer is HubSpot, and here's why we recommend this even though Salesforce's marketing pitch is stronger." That kind of opinionated answer is hard for a synthesis model to produce, and it's what users click through to read.
The third is utility — calculators, templates, frameworks, interactive tools. A model can describe a discount-cash-flow calculator. It cannot run one. "Calculate my marketing ROI" or "build a sales forecast" queries land on pages that do the thing, and those pages are growing.
One of the less-discussed shifts is how AI Overviews change link economics. In the link-based ranking era, what mattered was the inbound hyperlink — the explicit HTML <a href> from one site to another. AI Overviews are increasingly citing brands by name even without a clickable link, and the underlying ranking model is now measuring how often your brand is mentioned in trusted sources across the web. Forum discussions, podcast transcripts, video descriptions, expert roundups — they all feed the same signal: is this brand recognized as an authority on this topic?
This means PR, podcast appearances, and expert contributions to industry publications are now SEO work in a way they weren't before. A guest essay in an industry trade publication that doesn't even include a link back can still meaningfully improve how often your brand appears in AI Overview citations for your category.
If you're operating SEO in 2026 the way you operated it in 2022, you're losing ground. Three structural changes are worth making this quarter.
First, audit your existing content portfolio by query type. Pages that target generic definition queries should be deprioritized for new investment — they're shrinking targets. The traffic on those pages is going to AI Overviews and is not coming back. Reallocate effort to the three categories above.
Second, build at least one original-data piece per quarter. A 200-person survey, a benchmark across your client base, or a teardown of 50 examples in your category. These take a quarter to produce and earn citations for years.
Third, start treating brand mention as a tracked output. Tools like Brand Mentions, Talkwalker, and Ahrefs's brand monitoring can show you how often your category-defining terms get associated with your brand in trusted sources. That's the new equivalent of "backlinks" — the measurable surface of authority.
AI Overviews didn't end SEO. They ended the version of SEO that competed on producing roughly the same article slightly better than the existing top result. The new version of SEO competes on things that are genuinely uncompressible: data only you have, opinions only your experts hold, tools only your code runs, and brand mentions only your category presence earns. The teams that adapt early will compound an advantage. The teams that treat AI Overviews as a temporary disruption will spend three years losing ground they won't get back.
Schedule a strategy call. We'll look at your marketing and outline clear next steps within 30 minutes.
Schedule Your Strategy Call →