Is AI Overviews the Biggest AI Search Surface in 2026?

ChatGPT gets most of the AEO budget. Google AI Overviews now gets most of the queries. The gap has been widening for a year, and a 500,000-prompt study published this week makes it impossible to keep ignoring.

Peec AI's Tomek Rudzki published AI Overviews is the most undertracked AI search on May 28, 2026. The study pulled 500,000 business-intent prompts in April 2026 across six markets and eight query types. The headline numbers reset what anyone serious about AEO should be optimizing for.

AI Overviews appeared in 86.7% of US searches in April 2026, up from 56.9% twelve months earlier. That is 30 percentage points of expansion in a single year. The decision-stage finding is the one most teams will misread: 88.5% of commercial-intent queries return an AI Overview. The myth that AIO is informational-only is officially dead.

AI Overviews now sits between your buyer and your homepage on nearly nine of every ten commercial-intent queries.

This piece walks through the data, the diagnostic for why most AEO programs still under-track AIO, and the practitioner playbook for rebalancing your AEO budget toward the surface that reaches the most buyers.

Why AI Overviews is now the largest AI search surface

Google publishes the user counts. Peec AI publishes the appearance rate. Cross those two numbers and the conclusion is uncomfortable for anyone running a ChatGPT-first AEO program.

Reason #1: AI Overviews reaches 2.5 billion monthly users, roughly 3x ChatGPT

At Google I/O 2026 in May, Sundar Pichai disclosed an AI Mode 1B-user count and a 2.5B-user count for AI Overviews. ChatGPT's last confirmed user number was approximately 800M weekly active users from OpenAI's October 2025 update. We unpacked the I/O disclosure in Google AI Mode hit 1B users, queries got 3x longer. On a monthly-active basis, the ratio is roughly 3:1 in AIO's favor.

Reason #2: 86.7% appearance rate is up 30 percentage points year over year

Peec measured AIO appearance at 56.9% in April 2025 and 86.7% in April 2026. That is a 50% relative lift in twelve months. No other AI search surface has expanded its surface area at this pace inside a single year.

Reason #3: 88.5% commercial-intent appearance kills the "informational only" myth

For years the default AEO narrative held that AI Overviews triggers on questions and definitions and skips commercial queries. Peec's 88.5% finding on decision-stage queries inverts that assumption. The myth survives because most AEO dashboards undercount commercial AIO appearances. Vendor differences in query classification matter here.

Reason #4: Long-form queries trigger AIO at nearly 90%

Two-word queries trigger AIO 64.6% of the time. Eleven-to-fifteen-word queries trigger it 89.1% of the time. Google's disclosure that AI Mode queries are running roughly 3x longer than traditional Search means the share of long-form queries is growing across the board, and that growth flows directly into AIO appearance volume.

Reason #5: Almost every functional query type triggers AIO above 90%

By query category, AIO appears on short-fact, Boolean, instructions, definitions, reasons, comparisons, and consequences queries at 93 to 97%. The only category below 70% is "other" (vague or ambiguous queries) at 67%. If your buyers are asking real questions, AIO is in the result page.

Why most AEO programs still under-track AI Overviews

Brand teams are not stupid. The under-tracking has structural reasons, and naming them helps fix the budget conversation.

Reason #1: ChatGPT has the louder press cycle

Anthropic raised $65B at a $965B valuation on May 28. OpenAI shipped GPT-5.5 in April. Most AEO content marketing follows the model-release cycle, not the user-behavior cycle. Press coverage favors the surfaces that ship features weekly, and AIO ships in silence.

Reason #2: Vendor dashboards default to ChatGPT and Perplexity

The Tier-1 monitoring vendors all cover AIO, but most onboarding flows default to ChatGPT and Perplexity prompt sets. The default determines the chart you see at the next QBR. Defaults shape narratives, and the AIO chart often sits one tab away from the main screen.

Reason #3: AIO citations are harder to attribute back to traffic

Google Search Console still aggregates AIO impressions inside standard organic data. Until Bing Webmaster Tools shipped AI citation data, there was no surface giving real per-citation traffic attribution. AIO measurement requires inference, which loses budget arguments to surfaces that report direct numbers.

Reason #4: The France 0% finding makes AIO look like a US/UK story

Peec reports US 91.4%, UK 91.7%, Australia 86.5%, Germany 75.6%, Poland 82.9%, and France 0% appearance. France's 0% reflects regulatory exclusion under the press-publisher compensation regime, not a Google rollout gap. EU teams reasonably under-weight AIO when their largest market shows zero. The catch is that the EU regional average is 76%, not zero. Only France is excluded.

Reason #5: Conference demos render ChatGPT and Perplexity better

SEO conference talks lean on ChatGPT and Perplexity in their demos because the citation chips render cleanly on screen. AIO does not always show a sources box, and the answer surface varies by query. The surface that demos worst is the surface that gets the least mindshare, even when it has the most users.

Traditional AEO programs vs AIO-first programs

The structural difference shows up in how teams set priorities, not in what they think.

Traditional AEO programs ask:

• •Are we cited in ChatGPT for our top 30 prompts?
• •What does Perplexity say when a buyer asks about our category?
• •How do we lift Share of Model on Claude for B2B queries?
• •Which sources is GPT-5.5 pulling for our category?

AIO-first AEO programs ask:

• •Which commercial-intent buyer queries trigger an AIO, and are we in it?
• •Is our brand named inside the AIO response, or only in the sources list?
• •What does an 11-to-15-word commercial query about our category surface, and where are we in that answer?
• •How does our AIO citation rate compare to our ChatGPT citation rate on the same prompt set?

The first list is not wrong. The second list reaches a larger audience faster on the data Peec just published.

A 30-percentage-point year-on-year appearance gain is a budget signal, not a research footnote.

How to rebalance an AEO program toward AI Overviews

Below is the practitioner playbook. Five steps, each independently runnable, ordered by the sequence that produces a defensible baseline fastest.

Step 1: Audit citation share across all five AI surfaces, not just ChatGPT

Capture your brand's citation rate on ChatGPT, Claude, Perplexity, Gemini, and AI Overviews for the same prompt set. A typical B2B SaaS brand finds AIO citation rate is its weakest of the five surfaces, despite AIO having the highest user reach. That gap is the single biggest unmonetized AEO opportunity in 2026 for most mid-market brands.

Step 2: Segment prompts by intent stage and word count

Peec's data shows long-form, commercial-intent queries trigger AIO at the highest rate. Build a tracked prompt set that mirrors the way your buyers actually search: 11-to-15 words, comparison and decision-stage framing, vertical context. We covered the prompt-selection methodology in how to select prompts for LLM tracking.

Step 3: Rewrite top buyer pages for AIO passage extraction

The AIO ranker reads passages, not pages. Lead each major section with a 40-to-60-word direct answer. Put numbers and named studies in standalone sentences. We described the structural pattern in detail in passages beat pages.

Step 4: Add FAQ schema, but only where buyers genuinely ask the question

FAQ schema lifted AI citations 350% in Otterly's 1M-citation study. Selectively deploy it on comparison pages, pricing pages, and product-vs-product pages where buyers ask discrete questions. Skip it on category pages where the FAQ feels invented. Google's May 2026 core update is reportedly tightening on fake FAQ schema, so selectivity protects you on both surfaces.

Step 5: Track AIO appearance before tracking AIO citations

Most teams skip step five and go straight to "are we cited." If AIO does not appear on a tracked query, no amount of optimization will cite you. Build an AIO-appearance heatmap for your tracked prompt set first; then layer brand citation tracking on top. Without the appearance baseline, citation gains look bigger or smaller than they actually are.

What this means for your 2026 AEO budget

The Cite team is running internal client baselines with AIO as the primary surface for general-search-intent B2B SaaS workflows and ChatGPT as the primary surface for high-consideration product evaluation queries. A typical mid-market B2B SaaS client split now looks roughly like this.

Surface	Monthly users (approx)	Budget weighting today	Budget weighting after rebalance
AI Overviews	2.5B	15%	35%
ChatGPT	800M WAU	50%	30%
Gemini	900M	10%	15%
Perplexity	~22M WAU	15%	10%
Claude	~30M WAU	10%	10%

The numbers are illustrative, not prescriptive. Every client's actual split should come from real prompt-tracking baselines across their category, layered on the seven CITE metrics framework. What is consistent across our client baselines is the direction of travel: AIO budget weighting needs to rise, ChatGPT weighting needs to ease back, and the two need to be measured on the same prompt set so the comparison is honest.

What absorption looks like inside an AI Overview

Appearing in the AIO sources list is not the same as shaping the AIO answer. A page can be selected as a source and contribute nothing to the visible response. A page can also be paraphrased into the answer without showing up in the sources chip on mobile. The gap between selection and absorption is what we covered in how to measure AI citation absorption, and it applies to AIO with sharper teeth than other surfaces because the AIO source list is often abbreviated or hidden.

Three structural moves shift AIO absorption fastest:

• •Direct-answer paragraphs of 40 to 60 words sit immediately under each H2.
• •Numbers, percentages, and named studies appear in standalone sentences that can be quoted verbatim.
• •One-liner pull quotes summarize each section's claim in 12 to 20 words.

A page built this way wins the same number of AIO selection events as before, and does more answer-shaping work each time. The result reads in the dashboard as a citation rate that did not move and a Share of Voice that did.

FAQ

What is AI Overviews and how is it different from AI Mode?

AI Overviews is Google's AI-generated answer that appears at the top of a standard Google search results page. AI Mode is Google's full chat-style search experience that replaces the SERP entirely. AI Overviews reaches 2.5B monthly users and appears in 86.7% of US searches as of April 2026, per Peec AI's 500K-prompt study. AI Mode reaches 1B monthly users and shows roughly 3x longer average query length.

Why does Peec say France has 0% AI Overviews appearance?

France excludes AI Overviews under press-publisher compensation regulation tied to the Digital Services Act and adjacent French publisher law. The 0% figure reflects a regulatory exclusion, not a Google rollout gap. Other EU markets still see AIO at 75 to 83%: Germany 75.6%, Poland 82.9%. The EU regional average is 76%, not zero.

Is AI Overviews really larger than ChatGPT for buyer-facing queries?

On monthly active users, yes. AIO reaches roughly 2.5B people per month versus ChatGPT's ~800M weekly active users from the October 2025 OpenAI disclosure. On commercial-intent appearance, Peec measured AIO at 88.5%, which directly addresses the "AIO is informational only" claim. ChatGPT remains a critical AEO surface for product-discovery and high-consideration evaluation queries; AIO has the larger reach for general-search-intent workflows.

How should I split AEO budget between AI Overviews and ChatGPT in 2026?

Every client split should come from real prompt-level baselines, not category defaults. A reasonable practitioner reference for mid-market B2B SaaS is roughly 35% AIO, 30% ChatGPT, 15% Gemini, 10% Perplexity, 10% Claude. Adjust based on your buyer's actual query patterns and where your brand is currently being absorbed into answers.

What is the single biggest optimization for getting cited in AI Overviews?

Lead each major page section with a 40-to-60-word direct answer that matches the query phrasing. AIO reads passages, not pages, and the passage closest to the matching heading is the one that gets quoted. Pair this structural change with selective FAQ schema on commercial-intent pages and explicit named data in standalone sentences.

What to do this week

Run an AIO-appearance audit on your top 30 buyer prompts. If AIO appears on more than two-thirds of them and your brand is in fewer than a third of those answers, you have the budget rebalance the data is pointing to. If AIO appears on fewer than half of your tracked prompts, your prompt set is mis-calibrated for the surface that just became the biggest in AI search.

The Peec study reset the empirical anchor. The harder work is reframing what success looks like on AEO dashboards built for a surface that no longer has the most users. That reframe is the project for the rest of 2026.