10 SEO Trends I’ve Seen Firsthand in 2026 (With Data)

I work at Ahrefs, one of the world’s best-known SEO tools, which means I get to talk with and learn from SEOs every day. And spoiler: it is wild how much SEO has changed in the past year, and how much SEOs are now expected to change with it.

A brand new awareness and growth channel appeared in the form of AI search, and in most companies, it gets shoved into the SEO remit. There’s a whole new technical infrastructure to understand, too: it’s no longer enough to know how search engines work; you also have to know how AI search engines work.

Meanwhile, the classic SEO playbook of top-of-funnel informational content (as much as it pains me to admit) seems to work a little worse with every passing day. We are all in the experimental, throw-spaghetti-at-the-wall phase, trying to work out where the next era of compounding growth comes from.

These are some of the SEO trends I have watched shift over the past year. All of these trends are backed up by real Ahrefs data (to show these are genuine industry-wide trends, and not just something I’m imagining).

Ahrefs Brand Radar.

So a single question can quietly become a dozen searches behind the scenes, and the pages that get cited are the ones that answer those hidden queries. The same logic now shapes traditional ranking, too, because AI Mode sits on top of Google. If you want to show up, you need to rank for the head term and the fan-outs around it.

But there’s little use chasing exact fan-out queries: they are probabilistic, and change with every generation, so optimizing for a specific one is potentially a waste of time. Better: build a topic cluster that covers the whole topic space instead. Which is to say: the winning move is good old-fashioned keyword clustering and topical coverage, but pointed at a new search experience.

Keyword	Value	Change
query fan out	450/mo	+2,550% YoY
query fanout	200/mo	new (2025)
google query fan out	100/mo	new (2025)
query fan out seo	150/mo	new (2025)

Further reading

latest study found that AI Overviews reduce clicks to the top-ranking page by 58%. For every 100 clicks a number-one result would historically earn, Google now keeps 58 of them. We re-ran the analysis across 300,000 keywords using December 2025 Search Console data, and the effect held all the way down the page: position 2 lost 51%, position 3 lost 46%, and so on.

Then there is AI Mode, the longer-form, conversational version that feels a lot more like ChatGPT than a traditional search. Different format, same outcome. Across both, the story is the same: there are fewer clicks available in the SERP, your blue link is harder to find, and users are far more likely to get their answer on the page itself and never visit your site.

This goes well beyond a dip in traffic. It changes the unit economics of SEO, because the clicks you used to count on are quietly being kept by Google.

Keyword	Value	Change
ai mode	881,000/mo	+17% YoY
ai overviews	5,600/mo	+17% YoY
ai overview optimization	900/mo	+625% YoY
ai overview tracking	2,100/mo	+594% YoY

Further reading

what content engineering is, and why AI content is not inherently bad for SEO.

The data shows the attitude shift. Interest in creating with AI has climbed steadily since 2021, while interest in detecting AI content has fallen since its 2023 peak. People are normalizing the behavior. It is becoming more acceptable to use AI to produce high-quality work, which makes sense when you remember that every Google property, Docs, Sheets, Gmail, now has AI text generation built in.

Using AI to write is no longer a dirty secret. It is just part of how information gets made and shared on the internet.

Keyword	Value	Change
ai writing	19,000/mo	+15% YoY
ai content detector	14,000/mo	+11% YoY
ai content creation	5,100/mo	+16% YoY
humanize ai content	600/mo	+18% YoY

Further reading

semantic search and the knowledge graph if you want to go deeper.)

Keyword	Value	Change
semantic seo	2,800/mo	+13% YoY
topical authority	1,400/mo	+14% YoY
entity seo	700/mo	+15% YoY
semantic keywords	1,400/mo	+12% YoY

Further reading

Ahrefs MCP or Agent A, you can bring world-class SEO data into your AI workflows and automate things that simply could not be automated before.

We just published a list of SEO automations built with Agent A. The short version: keyword research, content updating, internal linking, site auditing, and a long tail of rote processes are all fair game now. The work that used to eat your afternoon is increasingly something you can hand to an agent.

Keyword	Value	Change
seo ai agent	900/mo	+285% YoY
ai seo agent	600/mo	+171% YoY
seo automation	2,600/mo	+14% YoY
agentic seo	150/mo	emerging

Further reading

llms.txt, and searches for it are up 154% year over year. I am including it here because, surprise surprise, it’s usually SEOs who get asked to implement it. But I want to be blunt about what the data says.

llms.txt is a proposed standard that no major AI search provider actually uses. As our research found, 28% of sites now publish an llms.txt file, and 97% of those files were never fetched at all in a given month. Of the small share that did get read, most of the requests came from bots that are not AI tools. Slackbot, the link-preview bot, fetched llms.txt files more often than PerplexityBot did. Google has said plainly that the file is not needed to appear in its generative AI features.

Bar chart shows ahrefs’ study of 137K domains. 97% of llm.txt files are never requested.

So at this point llms.txt delivers basically no benefit (and even carries a little risk, making it much easier for competitors crawl your sitemap). People like simple, concrete things they can do to influence mysterious processes like AI visibility, which is exactly why this one spread. If someone is pushing you to add one, send them this data. There are far better places to spend the hour.

Keyword	Value	Change
llms.txt	3,700/mo	+154% YoY
what is llms.txt	1,100/mo	+423% YoY
llm.txt	1,100/mo	+202% YoY
llms.txt generator	450/mo	+223% YoY

Further reading

research on ‘best’ lists found that of all the source types cited by ChatGPT and other AI search engines, this format was cited most often, showing up in 43.8% of ChatGPT’s sources and 48.9% of AI Overviews.

It makes sense. A lot of queries are commercial, and listicles are inherently commercial. They are genuinely useful in the buying process, as long as they are honest and not biased. But SEOs are going to SEOs: plenty of companies abused this by ranking themselves number one in ‘best’ lists published on their own domain. In Glen’s data, 67.6% of ‘best software’ SERPs featured a list where the author put themselves in the top spot.

Now the tactic may be waning. As Lily Ray has reported, recent Google updates have hit self-promotional listicles hard, especially in SaaS. It increasingly looks like Google may use these lists but credit the competitors named in them while disregarding the publisher’s own self-placement on its own domain.

Worth knowing if you are leaning on this play: it may not be as effective as it was, and honestly, I hope it fades out.

Metric	Value	Change
“best” lists in ChatGPT sources	43.8%	of all page types
prominence in AI Overviews	48.9%	highest of any platform
SERPs where author self-ranks #1	67.6%	250 “best software” SERPs
recently updated lists	79.1%	updated in 2025

Further reading

We analyzed 75,000 brands, and branded web mentions correlated with AI visibility at 0.66 to 0.71, with YouTube mentions edging even higher at around 0.74. The more your brand is talked about, in relevant contexts and across many places, the more likely it is to show up in ChatGPT, AI Mode, and AI Overviews.

This has revived a lot of interest in plain old brand building. It also lines up with traditional search. Branded search volume is very likely a signal Google uses, because whether real people search for your brand by name is a strong tell of whether a site is legitimate or just built for SEO. A site nobody searches for is easy to discount.

And this is the natural answer to the previous trend. If Google is pulling away from sites that only inform, the sites it still rewards are brands: businesses people seek out by name and that do something beyond publishing information.

So SEOs are increasingly trying to cultivate, grow, and protect their branded search. That means more than traditional SEO. It means brand marketing, advertising, investing in other formats, and generally doing the work to make your brand more visible across the internet.