SEO shifts toward authority and distribution as AI Overviews reshape search

Maintenance SEO is not growth SEO anymore

The safest SEO work is no longer the work that grows the business. Keyword research, on-page optimization, technical audits, content briefs, link building, and reporting still matter, but Claire Taylor’s argument is that they now function more like maintenance than a growth engine. They protect the baseline, keep pages clean, and stop obvious leaks; they do not, by themselves, guarantee new visibility in an AI-shaped search results page.

That shift matters because the old playbook was built for a world where publishing more pages, targeting more keywords, and tightening on-page relevance could usually buy incremental gains. Taylor’s point is that the market has changed underneath that model. When competent, generic content can be produced by anyone with a similar prompt, volume stops being a moat. If the page is interchangeable, ranking it is harder, and even when it ranks, it is often less valuable than it used to be.

AI Overviews changed the search contract

Google has already pushed AI Overviews into the mainstream. Sundar Pichai said in Alphabet’s Q1 2025 earnings update that the feature had more than 1.5 billion users per month, and Google later said AI Overviews were available in more than 200 countries and territories. Google also said the feature was driving more than 10% higher usage for the kinds of queries that trigger it in its biggest markets, including the United States and India.

That usage shift explains why SEO teams feel the ground moving. AI Overviews were launched in May 2024 and rolled out unevenly through 2025, which made the product feel volatile, but the broader trend is now hard to miss. seoClarity found AI Overviews on 30% of U.S. desktop keywords as of September 2025, with more than 99% of instances sourced from the top 10 web results. It also reported a 474.9% year-over-year increase on mobile, which is a brutal signal for anyone still assuming search results behave like they did a few years ago.

Ahrefs drew a similar conclusion from a different angle. In its study, AI Overviews reduced clicks by 34.5%, appeared on 9.46% of desktop keywords in its dataset, and showed up on 12.8% or more of all Google searches by volume. The important detail is not just that AI Overviews appear often, but that they compress the click path. If the answer is already visible, the incentive to visit the source drops fast.

The biggest hit is at the top of the funnel

The clearest damage shows up in informational searches, where classic SEO has long depended on broad keyword coverage. WebFX found AI Overviews in 25.8% of 2.37 million U.S. queries, and more than 50% of queries with seven words or longer. That matters because longer queries are often the ones where searchers are asking real questions, comparing options, or looking for a first pass at a decision.

WebFX also found especially high prevalence in health, finance, jobs and education, and family-related searches. Those are exactly the categories where users want quick synthesis, and exactly the categories where a zero-click answer can siphon away the first visit. Semrush’s 2025 data shows the pattern is broadening too: in January 2025, 91.3% of queries triggering AI Overviews were informational, but by October that share had fallen to 57.1%. Navigational queries rose from 0.74% to 10.33% in the same period, which means AI Overviews are no longer confined to the early research stage.

That matters because high-volume keyword lists depend on stable intent patterns. If the system that answers the query keeps changing, the list gets less predictive. Taylor’s broader warning is that chasing volume in a search environment dominated by AI summaries can create a lot of work without creating a lot of growth.

What actually moves visibility now

The newer work is not mysterious, but it is harder to fake. Visibility now depends on whether the brand shows up as something worth citing, mentioning, searching for, and trusting across the places search engines and AI systems pull evidence from. That means authority, distribution, and brand demand are doing more of the heavy lifting than a spreadsheet of target terms ever could.

In practice, the strongest SEO programs now look less like content factories and more like business development for organic discovery. The pages that win are usually anchored in one of three things: proprietary information, product reality, or unmistakable brand signal. Generic explainers are cheap to produce, but unique data, first-party research, product-led tutorials, and clear expert positioning are still expensive to copy.

A useful reset looks like this:

• Build pages around data you own, not just topics everyone else can summarize.

• Write product-led content that shows how the thing works in the real world, including setups, edge cases, and tradeoffs.

• Turn brand demand into a search asset, because people looking specifically for you are harder for AI summaries to intercept.

• Seed visibility across formats and platforms, not only on the site, so the brand appears wherever discovery starts.

• Measure downstream outcomes, not just rankings, because a position that never earns the click is a vanity metric in a zero-click environment.

That last point is not new, but Taylor’s earlier Search Engine Land argument makes it sharper: ranking reports and position tracking do not fully capture SEO value in an AI-heavy, zero-click world. The key question is no longer simply whether you got discovered. It is whether that discovery translated into attention, trust, and eventually revenue.

The new job of SEO is broader than retrieval

This is the real reset. Traditional SEO was built to answer a retrieval problem: match a query, earn a rank, collect a click. AI Overviews have made that chain less reliable, because the answer can now be synthesized before the visit ever happens. In that world, the work that matters most is not just optimizing for retrieval, but building the evidence that makes a brand visible enough to be cited and memorable enough to be searched directly.

That is why maintenance SEO still has a place, but only as the floor. The ceiling now comes from authority, distribution, and proprietary value. If your program still treats SEO as a keyword engine, you will keep defending your base and wondering why growth stalled. If you treat it as a broader business discipline, you get a better shot at the searches that still lead somewhere.