Answer Engine Optimization Rises as AI Overviews Cut Clicks

Ranking first no longer means getting the click

For twenty years, search marketing worked on a simple trade. Publishers wrote pages, Google ranked them, and users clicked through. By late 2026 that trade looks broken. Google's AI layers now answer a large share of queries directly on the results page, and a growing industry argues that the goal has moved from ranking to being cited. The name for this work, answer engine optimization or AEO, has spread quickly through marketing playbooks.

The term is not just hype, but it needs some care. The evidence that AI summaries reduce clicks to publishers is now strong. The evidence that AEO reliably wins that traffic back is much thinner, and a lot of it comes from companies that sell AEO tools.

The evidence on lost clicks has caught up with publishers' complaints

Google launched AI Overviews for all U.S. users in May 2024, and publishers started reporting traffic losses soon after1. At first, Google could treat these complaints as anecdotes. Head of Search Liz Reid argued in August 2025 that total organic clicks had stayed "relatively stable" and that click quality had improved slightly7. Since then, the data has piled up against that view.

Pew Research Center tracked real browsing behavior. Users clicked a traditional result 8% of the time when an AI summary appeared, compared with 15% when it did not. Only 1% clicked a link inside the summary itself5. Ahrefs first measured a 34.5% drop in position-one click-through rate in April 2025. When it reran the study on December 2025 data, the drop had grown to about 58%, which Ahrefs described as Google "keeping" 58 of every 100 clicks a top page used to earn6.

The strongest finding so far is a randomized field experiment by Saharsh Agarwal of the Indian School of Business and Ananya Sen of Carnegie Mellon. They used a browser extension to show or hide AI Overviews for 1,065 desktop Chrome users. When an Overview appeared, outbound organic clicks fell 39.8% and zero-click searches rose 34.5%. Sponsored clicks did not change2. Removing the Overview did not reduce engagement on the sites users reached, which challenges Google's argument that the remaining clicks are simply better quality2. The effect was concentrated where Overviews usually sit, at the very top of the page, and it was driven mostly by informational queries9.

The size of the losses depends on who is counting. Amsive's study of 700,000 keywords found an average click-through decline of only 15.49%. Branded queries actually gained about 18%8. At the other extreme, DMG Media told the UK's Competition and Markets Authority that the Daily Mail's desktop click-through rate fell 89% when an Overview appeared above its link1.

The gap between these figures reflects real differences in query mix and method, not just noise. The most defensible reading is this: informational, non-branded content loses heavily, while brand searches are mostly spared. The randomized estimate of roughly 40% is the most reliable single number available.

Not every drop is caused by AI Overviews

Two caveats complicate the picture. First, the same Ahrefs study that produced the 58% figure also found that position-one click-through on informational queries without an AI Overview fell from 7.6% to 3.9% over the same period10. Seer Interactive found a similar pattern: non-Overview queries lost 41% of organic click-through9. Search behavior was changing beyond Overviews alone, so tying every lost visit to them overstates the case.

Second, the trend may not run in only one direction. Seer data from April 2026 shows organic click-through on Overview queries falling to 0.61% in September 2025 and then recovering to 2.4% by February 20264. That rebound coincides with Google's efforts to push traffic back out, so treating the decline as permanent may be premature.

The damage is also uneven. Chartbeat data shows sites with 1,000 to 10,000 daily page views lost about 60% of their Google referrals over two years, compared with 22% for large publishers10. Press Gazette's analysis put the drop in global Google referrals to major publishers at about 33% in the year to November 20254. Small and niche publishers, whose specialized answers are easiest to summarize, are taking the worst of it.

Google is changing ranking and the results page at once

Google has kept up the pressure. At I/O 2026, it said AI Mode had passed one billion monthly users, AI Overviews had reached 2.5 billion, and the search box was getting its biggest redesign in more than 25 years30. Two days later, on May 21, Google began rolling out its second broad core update of the year21. That update shipped without a companion blog post explaining its goals24.

Because the ranking change and the interface change landed together, it became hard to diagnose traffic changes. A site could keep its rankings and still lose clicks because AI features answered the query above it21. Practitioners now point to a common Search Console pattern: steady impressions but falling clicks. They read it as a sign that the problem happens before the click, not that rankings dropped23. Google's own figures show the same split, with impressions up 49% and clicks down 30%10.

Reports differ on how far the shift has gone. One account says AI Mode is now the default search experience worldwide27. Another says it is the default "for many queries"26. A third still describes AI Mode as a separate experience from inline Overviews28. Coverage agrees that Google is merging the two surfaces into one continuous flow30. The safest conclusion is that this merger is still underway rather than finished.

Estimates of how often AI answers appear also vary. Figures from one dataset suggest 48% of searches showed an AI answer by March 202621. The randomized study observed Overviews on about 41% of queries2. A study cited by Contently put the rate closer to 25% in early 202614. These numbers depend on how queries are sampled, but all of them show AI answers have become a routine part of search.

What AEO actually asks publishers to do

Against this backdrop, AEO has become the default advice. Most guides describe it the same way. Lead each section with a direct answer of 40 to 60 words, use a clean heading structure, add schema that matches what is visible on the page, and include original data so that AI systems have a reason to cite you1215. The goal changes from ranking position and click-through rate to how often a brand is cited and how much of the AI answer it occupies12.

The strongest argument for this approach is a growing gap between rankings and citations. One analysis says top-10 pages supplied 76% of AI Overview citations in mid-2025 but only about 38% by early 202628. Others put the overlap between top rankings and AI Mode sources at 17% to 36%27. Pages that are cited inside an Overview reportedly get about 35% more organic clicks than pages that are not27. If those figures hold, being cited has its own value, separate from ranking.

However, the guides contradict their own headlines. Almost every one says AEO does not replace SEO. Pages still need to be crawlable, indexed and authoritative before any AI system will pull from them1318. Google's own guidance says AI Overviews and AI Mode draw on the same index and quality signals as organic results, and it dismisses llms.txt files as unnecessary for Google Search24. Analysts tracking Google's January crackdown found that sites losing organic visibility also tended to lose presence in AI features23. Google has also confirmed that its spam policies now cover attempts to manipulate or buy AI citations29.

The best reading is that AEO is SEO with a different scorecard, not a separate discipline that makes the old one obsolete. Writing answer-first and structuring pages clearly is reasonable editorial practice. The specific statistics around it, such as a 2.8x citation lift for nested headings12, come largely from vendors selling AEO software and should be treated as marketing claims until independent researchers test them.

A growing legal and regulatory fight

Technique will not fix the economics. Penske Media's antitrust suit describes Google turning from a search engine into an "answer engine" that keeps users on its own pages5. The European Publishers Council has filed a complaint with the European Commission3. In May, Google announced five changes meant to send traffic back to publishers, including a "Further Exploration" links section and subscription labels3. Meanwhile, Search revenue rose 17% year over year to $63.07 billion in Q4 2025, and ads alongside AI Overviews expanded sharply4.

That contrast between Google's growing revenue and publishers' shrinking traffic is the core of the story. AEO can help an individual site win a larger share of a shrinking pool of clicks. It cannot make that pool bigger. Whether publishers recover their traffic will depend less on how they format answers than on regulators, courts, and whether Google's new link features produce measurable gains.