If your search pipeline depends on “best X” and “X vs Y” pages, the constraint is simple: AI Overviews don’t show up evenly across query types, so your “AIO rate” might be measuring the wrong thing.
One dataset can tell you AI Overviews are everywhere. Another can tell you they’re still a minority feature. Both can be true—because the sample definition is doing more work than most dashboards admit.
That sounds academic. It isn’t. If the board asks why organic-assisted pipeline is wobbling while rankings look “fine,” this is one of the few SERP changes that can explain the disconnect without hand-waving.
The data isn’t inconsistent. The samples are.
Search Engine Journal reported on a Peec AI analysis that found Google’s AI Overviews appeared in about 87% of 500,000 prompts in its dataset, which skewed toward commercial, buying-intent queries and excluded navigational searches. In that same analysis, decision-stage prompts triggered AI Overviews at 88.5%. (Source: Search Engine Journal / Peec AI analysis)
Read that again: this isn’t “AI Overviews appear on 87% of Google searches.” It’s “AI Overviews appear on 87% of this particular prompt set,” and the prompt set is exactly the kind of query mix B2B marketers care about when they’re trying to create qualified pipeline from non-branded demand.
But broader studies that include more query types often show much lower prevalence. SellersCommerce cites a study where commercial keywords were 6–8% of AI Overview triggers. BrightEdge reported AI Overviews on 23% of e-commerce queries (important, but not the same thing as “all commercial intent”). (Sources: SellersCommerce; BrightEdge)
So which number should a CMO believe? The better question: what exactly is being measured? Query mix. Intent taxonomy. Region. Even rollout status. Those choices change the story.
Why this matters now: AI Overviews are moving down-funnel
AI Overviews weren’t broadly launched until 2024, so older “trend” charts can be noise if they treat pre-rollout SERPs as comparable. (Source: rollout timing note in research brief)
In 2025 tracking covered by Search Engine Land, the share of AI Overview appearances attributed to commercial queries increased from 8% to 18%, and the transactional share rose from 2% to 14%. (Source: Search Engine Land tracking)
That’s the part that should hit the RevOps nerve. As AI Overviews expand from mostly informational queries into commercial and transactional territory, they start intercepting the pages that used to do quiet, compounding work: comparison pages, “best for” pages, implementation checklists, pricing explainers.
And there’s a second-order effect that’s easy to miss: SEO teams can “win” visibility while losing clicks. Multiple industry perspectives flagged the risk that AI Overviews reduce click-through by answering queries directly on the SERP. (Sources: Search Engine Journal; Search Engine Land; agency commentary summarized in the research brief)
One move for this week: rebuild your AIO baseline around intent
If you only change one thing, change this: stop reporting a single “AI Overview presence rate” across your whole keyword set. It blurs the exact segment where the business impact is likely to show up.
Here’s the 5-minute version you can run this week: treat AI Overviews like a new SERP layer that varies by intent, then measure it the same way you’d measure any other channel shift—by segmented baselines and directional attribution, not vibes.
Hypothesis (make it falsifiable)
If we segment our priority non-branded keyword set into informational vs commercial vs transactional intent and track AI Overview presence separately, then we’ll see a higher AIO presence rate in decision-stage/comparison terms than in pure transactional terms, because industry tracking shows AIO expanding into commercial intent unevenly and some analyses suggest lower AIO frequency on direct product/service queries. (Sources: Search Engine Land tracking; Zapier analysis summarized in the research brief; SEJ/Peec AI decision-stage rate)
Setup / Launch / Readout / Next test
- Setup: Pull 200–500 non-branded queries that historically assisted pipeline (not just drove sessions). Include “best,” “vs,” “alternatives,” “pricing,” “reviews,” and “implementation” patterns. Owner: SEO lead + RevOps analyst.
- Intent tagging: Use clear rules before tools. Example: “vs/alternatives/best” = commercial (decision-stage), “pricing” = transactional-ish (but check the SERP), “how to/what is” = informational. (Directional, not definitive.)
- SERP capture: For each query, record: AI Overview present (Y/N), top cited domains if visible, ad presence, and your ranking. Keep it simple. A spreadsheet works; a SERP monitoring tool is fine if it reliably detects AI Overviews.
- Readout: Compare AIO presence by intent bucket. Then overlay Search Console: impressions, clicks, CTR, and brand vs non-brand splits for the pages mapped to those queries.
- Next test: For the bucket with the biggest AIO presence, pick 5–10 pages and rewrite intros to answer the comparison question faster, add explicit selection criteria sections, and tighten on-page entity clarity. Then re-check whether impressions rise while CTR falls (classic “on-SERP visibility up, off-SERP traffic down” pattern agencies have warned about). (Source: agency perspective in research brief)
Success metrics and guardrails
- Primary metric: segmented CTR change for the decision-stage bucket (not sitewide CTR).
- Secondary metrics: impressions (visibility), assisted conversions / pipeline influence (directional), and paid search CPC/CPA movement on overlapping terms.
- Stop-loss: if decision-stage clicks drop materially while impressions rise and pipeline influence declines for two consecutive reporting cycles, pause content expansion in that bucket and shift effort to pages with clearer conversion capture (demo/pricing/BOFU), plus paid coverage where appropriate.
The trade-off is real: this approach can reduce reported “SEO wins” in the short term because it forces uncomfortable segmentation. It also makes it harder to hide behind blended averages. Good.
Paid doesn’t disappear. It gets rearranged.
There’s another reason commercial-query AI Overviews matter: monetization. Google’s help documentation indicates that ads can appear alongside AI Overviews when Google detects commercial intent and relevant ads are available. (Source: Google help documentation referenced in research brief)
So the practical read is not “SEO is dead” or “PPC is dead.” It’s that the SERP is becoming a denser page where organic clicks are a scarcer resource, and paid units can sit adjacent to AI-generated answers. That changes creative fatigue dynamics, it changes what “above the fold” means, and it makes SEO/PPC handoff a measurement problem, not a channel-turf problem.
Kicker: the baseline you pick is the story you tell
The most dangerous thing about AI Overviews in 2026 isn’t that they exist. It’s that they can look rare or ubiquitous depending on what someone put in the dataset.
Peec AI’s 87% number and the lower, broader-market numbers don’t cancel each other out. They’re a warning label: if commercial intent is where qualified pipeline gets created, then commercial-intent SERP measurement is where the reporting baseline needs to live.
