Ahrefs analyst Louise Linehan spends her days watching what people search for before they search for anything else — the meta-queries, the terms marketers type when they're trying to figure out if any of this is working. On July 24, 2026, she published a rundown of five ai search trends she's watching heading into the back half of 2026, built on Ahrefs' own keyword-volume data rather than vibes or vendor surveys. Two of those trends, read together, tell you more about where enterprise AI search budgets are actually stuck than any single number does. Two more, read together, describe a collision most teams haven't spotted yet: the same instinct pushing companies to over-produce machine-readable content for AI agents is the instinct that just got a wave of programmatic SEO sites penalized.
The AI Search Trends Ahrefs Is Tracking Right Now
Louise Linehan's title undersells the piece a little: "5 AI Search Trends I'm Seeing in 2026, Backed by Ahrefs Data" reads like a light trend roundup, but the underlying data is a real signal of where the market's attention is moving. Ahrefs pulled it from its own keyword database — actual search volume on the terms marketers and in-house SEOs are typing into Google, not survey responses or vendor self-reporting. That distinction matters. When "ai rank tracking" search volume jumps 175% and "agentic seo" jumps nearly 6,000%, that isn't a vendor's marketing claim about the state of the market. That's tens of thousands of practitioners changing what they search for, in real time, because their job changed under them.
We've been tracking this data series since Ahrefs' first pass at it earlier in 2026, when we broke down an ai search content strategy built on a billion Ahrefs data points. Linehan's July update is the sequel, and it's more useful, because it names five specific fault lines instead of one broad shift. Read Ahrefs' trend analysis if you want the raw percentages straight from the source. Read on if you want to know which two of these trends are actually the same trend, and which two are about to collide.
Trend 1: ROI Reporting Demand Is Outrunning the Data
Marketers want proof, and they want it in a format finance will accept. Ahrefs clocked "ai search tracking" search volume up 184% and "ai rank tracking" up 175%, which is a blunt but honest signal: people who used to track rankings are now trying to track citations, and they're doing it because someone above them is asking for numbers. Linehan points to one data point that captures the moment. Google's AI performance report inside Search Console — the panel showing how your pages show up in AI Overviews and AI Mode — became, in Ahrefs' own words:
“the most requested report in Google Search Console ever”
That's a genuine milestone: an entire industry pointed a data request at Google at the same time, and Google built the thing. The catch, per the practitioner feedback Linehan cites, is that the report still doesn't answer the question that actually matters to a CFO. It shows impressions and appearances. It does not reliably show whether a citation turned into a visit, a signup, or a dollar. The open question sitting underneath all that new search volume is simple and uncomfortable: are AI citations converting to real business outcomes, or is "we got cited" becoming a vanity metric with a shinier dashboard than the one it replaced. Nobody has a clean answer yet, which is exactly why the search volume keeps climbing instead of settling.
Search volume growth behind the demand for proof (Ahrefs, published July 24, 2026)
Trend 2: Agent Optimization Becomes a Real SEO Audit Line Item
The second trend is about who's reading your content at all. In June 2026, Cloudflare CEO Matthew Prince said agentic — bot — traffic had crossed 50% of all internet traffic for the first time, meaning more requests hitting the open web now come from automated agents than from humans clicking around with a mouse. That's the backdrop for the search terms Linehan flagged: "agentic seo" searches grew 5,867% and the broader optimization search cluster grew 852%. Those aren't typos. They're a signal that an entire practice area — optimizing so autonomous agents, not just crawlers or chatbots, can find, parse, and act on your content — went from a niche conversation to a line item almost overnight.
Google didn't wait for the terminology to settle before it acted. It added agentic-readiness checks to Chrome Lighthouse, which means AI-agent accessibility now sits next to Core Web Vitals and classic SEO checks as something Google audits by default. That's a meaningful tell. Lighthouse doesn't add categories for fads. If your technical SEO process still treats "can an AI agent complete a task on this page" as an edge case instead of a checklist item, you're already behind the tool that's supposed to be checking your homework.
Trend 3: The Traffic-Loss Conversation Moves From Panic to Action
The third trend is the one closest to home for anyone who's already read our research on the gap between citations and clicks: AI Overviews cut clicks to the #1-ranked organic result by roughly 58%. That number isn't new to regular readers of this site, and Linehan's piece corroborates it rather than revising it. What's changed since we first covered it is the tone of the conversation. Teams have stopped arguing about whether AI Overviews cost them traffic and started asking what to actually do about it.
Regulation is doing part of the work here. The UK's Competition and Markets Authority has real requirements bearing down on how Google surfaces AI answers, and a Munich court ruling went further, holding Google liable for the accuracy of its AI-generated answers rather than treating them as a neutral pass-through. Put those two pressures next to the click-loss number and you get a market finally moving past complaint and into strategy: opt-out decisions, recovery plans, and — for a lot of teams we work with — a hard look at whether it's time to refresh older content specifically so it survives AI citation extraction instead of just ranking well on a results page fewer people click through from at all.
Trend 4: GEO, AEO, and SEO Still Don't Have a Shared Name
The fourth trend is a naming problem hiding inside a strategy problem. Ahrefs found "generative engine optimization" search volume up 997% over 18 months, and "geo vs seo" comparison searches up 982% over the same window. Neither number is trivial, and together they describe an industry that still hasn't agreed on what to call the work everyone's suddenly doing more of. The debate splits into two camps: some argue this is a genuinely new discipline with its own techniques and its own success metrics; others insist it's traditional SEO wearing a new badge because the old one stopped testing well in a boardroom.
| SEARCH TERM | 18-MONTH GROWTH | WHAT IT SIGNALS |
|---|---|---|
| "generative engine optimization" | +997% | New-discipline framing gaining ground |
| "geo vs seo" comparisons | +982% | Practitioners still sorting old from new |
| "agentic seo" | +5,867% | Agent optimization entering the vocabulary |
| Broader optimization search cluster | +852% | Demand outrunning agreed terminology |
We don't think this naming debate is cosmetic, and we'll say where we land on it further down. For now, the practical read is simpler: whatever you call it, buyers are searching for it under at least four different labels, and your own content needs to answer to all of them, not just the one your internal team prefers.
Trend 5: Programmatic SEO's Cautious Comeback Meets the Agent-Optimization Rush
The fifth trend is where this gets genuinely risky. Interest in scaled, programmatic content creation rebounded 124% over 18 months, and the reason is obvious: AI made mass production of pages cheap and fast in a way it wasn't two years ago. Spin up a database, template a page, generate a few thousand variants, ship them. For comparison and alternatives content specifically, that approach can hold up — a well-built comparison page answering a genuinely different query for each entry is legitimate content, not filler.
But Google's March and June 2026 spam updates already punished a lot of sites that pushed this too far. Linehan's data shows a split: some tool-page-style programmatic sites are holding their traffic fine, because the pages behind them still answer distinct, real queries. Others scaled to thousands of thin, near-duplicate pages and got flattened across the same two update cycles. The difference wasn't the tactic. It was whether the tactic produced anything a human or an AI agent actually needed to read.
Here's where we think the real story is, and it's one Linehan's piece sets up without quite naming it: trend five and trend two are on a collision course. Teams chasing agent optimization are being told, correctly, to make content more structured, more extractable, more machine-readable — the same instincts that just got programmatic SEO sites penalized when they scaled without adding real value underneath the structure. More machine-readable content, faster, sounds like an obviously safe move when the goal is agent accessibility. It isn't. It's the exact shape of the mistake Google's spam updates were built to catch. A thousand thin pages templated for a crawler and a thousand thin pages templated for an AI agent fail an update for the same reason: nothing distinct behind the structure. This matters most for B2B SaaS teams running comparison or tool-directory content at scale, because that's exactly the content shape most tempting to template and multiply.
Why These AI Search Trends Are Really One Problem
Step back from all five and two of them stop looking like separate trends. Trend one is marketers demanding proof that AI citations convert to revenue. Trend four is the industry failing to agree on what to even call the work that's supposed to produce those citations. Those aren't adjacent problems. They're the same problem wearing two faces. You can't build a defensible ROI case for a practice when half your team calls it GEO, a third calls it AEO, and the rest insist it's just SEO with new vocabulary. Finance doesn't fund categories it can't name and can't audit, and right now the ai search trends Ahrefs is tracking are asking budget owners to do both at once.
That's the real reason enterprise budget commitment is lagging the search-volume growth in Ahrefs' data. It isn't a technical problem. Crawlers can be unblocked, schema can be added, an llms.txt file can ship in an afternoon. The stalled part is organizational: nobody's signed off on a name for the line item, and nobody's produced a reporting model a CFO trusts enough to fund at scale. Fix the naming and the measurement together — pick a term, define what "working" means in numbers finance already trusts, like assisted pipeline or branded search lift, not citation count alone — and the budget conversation moves. If you're building that case internally, our content marketing team has been building the reporting layer alongside the content for exactly this reason: a citation nobody can tie to revenue doesn't survive a budget review.
What To Do About These AI Search Trends
Don't wait for the industry to agree on a name. Pick one internally this quarter — we use GEO paired with classic SEO reporting — and build a single dashboard that ties citation rate to at least one number finance already trusts: assisted pipeline, branded search lift, or demo requests from AI-referred sessions. Then run one audit before you scale anything for agent readability: pull your last 90 days of published pages and check whether each one would survive a manual review for distinct value, not just structure. If a chunk of your roadmap is templated pages built to be agent-readable rather than reader-useful, that's exposure to the next spam update, not a competitive edge. Do the naming and the audit in the same month. Everything else in this piece is downstream of getting those two things settled.
See where you are cited today
A free snapshot audit of your rankings and AI citations before we ever talk.
Josh leads work at the intersection of SEO and generative engines at Something Inc., helping B2B brands get ranked and cited across every major AI engine.