If lower CTR pushes you to move budget between campaign types, check invalid traffic risk by Google Ads campaign type first, because the risk is not the same everywhere.
This post covers three things: what independent data shows about AI Overviews and paid clicks, what Google says about ads in AI Overviews and AI Mode, and how bad traffic actually affects your bidding. It ends with a checklist you can run this month.
What the data shows about AI Overviews and paid clicks
On queries with an AI Overview, paid CTR has been lower than on queries without one, and the gap has been narrowing. Organic results take the bigger hit. The three studies below use different methods, so read them as direction, not as one number to forecast with.
| Study | Sample | Finding |
|---|---|---|
| Seer Interactive, April 2026 | 53 brands, 5.47M queries, 296.9M paid impressions, Jan 2025 to Feb 2026 | Paid CTR with an AI Overview rose from 14.64% to 16.21%. Without one, it fell from 25.98% to 21.85% |
| Seer Interactive, September 2025 update (published November 2025) | Informational queries, 3,119 terms, 42 clients, Jun 2024 to Sep 2025 | Paid CTR on AI Overview queries fell from 19.70% to 6.34% (down 68%); on other queries from 19.1% to 13.04% (down 32%) |
| Pew Research Center, July 2025 | Browsing data from 900 US adults, March 2025 | Users clicked a traditional result in 8% of visits with an AI summary vs 15% without; 1% clicked a link inside the summary |
Sources: Seer Interactive, AIO Impact on Google CTR: 2026 Update, Seer Interactive, September 2025 update (published November 2025), Pew Research Center.
What to take from this:
- The two Seer studies measure different cohorts. Seer's 2025 data showed a sharp fall in paid CTR on informational AI Overview queries. Seer's 2026 study uses a broader, all-intent cohort, so it isn't directly comparable with the 2025 informational-query study. In that cohort, paid CTR on AI Overview queries edged up from 14.64% (Jan 2025) to 16.21% (Feb 2026). Seer offers a possible explanation for the 2025 drop, a layout change that "pushed ads further down on AIO-heavy queries," but says the cause is not known.
- Correlation, not cause. Seer's 2026 study notes: "We cannot claim causation. Higher-authority brands are also more likely to be cited." The paid sample in the 2025 study was much smaller than the organic one.
- People click less overall when a summary appears. Pew found that users ended their browsing session on 26% of pages with an AI summary, against 16% of pages with traditional results only. Pew measured organic links, but the same attention shift may also affect ads on those pages.
Where ads can appear: Google's rules for AI Overviews and AI Mode
Google decides whether your ad appears with an AI Overview, and you can't target or exclude that placement. Its About ads and AI Overviews help page sets out the rules:
| Question | Google's answer |
|---|---|
| Where can ads show? | Above or below AI Overviews in the 200+ markets where AI Overviews are available, and within AI Overviews in some markets |
| Which ads can show within AI Overviews? | Text and Shopping ads from existing Search, Shopping and Performance Max campaigns |
| Which markets for ads within AI Overviews? | English on mobile and desktop in Australia, Canada, India, Indonesia, Kenya, Malaysia, New Zealand, Nigeria, Pakistan, Philippines, Singapore and the US |
| Any excluded categories? | No ads within AI Overviews for sensitive verticals such as adult, alcohol, gambling, finance, healthcare and politics |
| Is there separate reporting? | No: “Google Ads currently doesn’t offer segmented reporting when ads show within Search AI Overviews” |
| Can I target or opt out? | No: “you can’t opt out of serving ads in AI Overviews” |
AI Mode is Google's conversational search experience. When Google announced it was testing ads there in May 2025, it said advertisers using "Performance Max, Shopping and Search campaigns with broad match, including AI Max for Search campaigns," would be eligible for AI Overviews and AI Mode (Google Ads & Commerce Blog, 2025). In May 2026 Google said it was testing two new ad formats in AI Mode, Conversational Discovery ads and Highlighted Answers (Google Ads & Commerce Blog, May 20, 2026). Eligibility for ads within AI Overviews is tied to broad match, AI Max, Performance Max, Shopping or Dynamic Search Ads (Google Ads Help: About ads and AI Overviews), and Google has been moving Search campaigns toward AI Max (see our post on the Dynamic Search Ads move to AI Max). That means more of your budget can end up on AI surfaces without a separate line in your reports.
The practical result: you have to infer the AI Overview effect from query-level data, because Google Ads won't show it to you directly.
How bad traffic actually hurts your bidding
Smart Bidding learns from conversions, not from analytics sessions. Google describes Smart Bidding as strategies that "optimize for conversions or conversion value in every auction," and notes that "you need to have conversion tracking enabled" (Google Ads Help: About Smart Bidding). At auction time it uses signals such as device, location, time, browser and the search query.
That changes which kind of bad traffic matters:
| Bad traffic type | Cost | Effect on Smart Bidding |
|---|---|---|
| Invalid click Google filters | Not charged, according to Google | Removed from your metrics |
| Bot click Google misses, no conversion | You pay for the click | No conversion to learn from. It can lower the measured conversion rate of a segment, but it doesn't make bidding seek out more bots |
| Bot or spam conversion (form fill, call) | You pay for the click and the follow-up time | Counted as a success, so bidding looks for more traffic like it |
| Real person, junk lead counted as a conversion | You pay for the click and the sales time | Same as above: bidding treats it as a win |
| GA4 event imported as a primary conversion (for example, an engaged session) | Depends on the event | Any bot that triggers the event becomes a conversion |
So the damaging path is not "AI crawlers create sessions that Smart Bidding learns from." It is fake or low-quality conversions being counted as primary conversions. Our guide to how fake leads train Smart Bidding covers the fixes in detail, and how invalid traffic damages lead quality covers the wider pattern.
Is automated traffic rising? On the web in general, yes. The 2026 Thales (Imperva) Bad Bot Report found bots made up 53% of internet traffic, with 40% classified as bad bots (Thales, 2026 Bad Bot Report). That figure covers all web traffic, not paid clicks. A crawler fetching your page to build an AI answer doesn't create a paid click unless it clicks an ad, and we know of no published data showing that AI Overviews increase invalid clicks on ads. AI agents that browse and click on behalf of users are a separate question, covered in our post on AI agent clicks and invalid traffic.
Which accounts are most exposed
Exposure depends on how much of your spend goes to informational queries and how clean your conversion data is.
| Account type | Main exposure | Why |
|---|---|---|
| Broad match or AI Max heavy | Query mix shifts, unseen AI surface spend | Broader matching reaches more informational queries, where AI Overviews appear most |
| Performance Max | Limited visibility | Eligible for AI Overviews and AI Mode, with no separate reporting for those placements |
| Lead generation | Bidding toward fake leads | If form fills count as primary conversions, any bot or junk lead teaches bidding the wrong thing |
| Finance, healthcare and other sensitive verticals | Less in-panel ad exposure | Google does not show ads within AI Overviews for these verticals, though ads can still show above or below |
| Exact and phrase match on transactional terms | Lower | Less exposure to the informational queries where AI Overviews appear most |
Checklist: what to check and change this month
- Split search terms into informational (how, what, why, vs, review) and transactional (price, quote, buy, near me) and compare CTR and conversion rate over the last 12 months.
- Compare Search impression share with Search lost IS (rank) and lost IS (budget), to tell a bid or budget problem from a shrinking query pool.
- List every conversion action and its status. Raw form fills, page views and GA4 engagement events should be secondary, not primary.
- Import qualified or closed leads as the primary conversion through offline conversion import or enhanced conversions for leads.
- Retract conversions you know are fake using conversion adjustments. Google Ads does not accept negative conversion values.
- Add negatives for informational queries that never lead to sales in your CRM data.
- Add the Invalid clicks column and check the Invalid Activity Credit Report for Search and Performance Max.
- Keep click logs for at least 60 days so you have evidence within Google's investigation window.
Adjustments that help:
- Split campaigns by intent. Keep transactional terms in their own campaigns, so a fall in informational CTR doesn't hide inside a blended average.
- Judge AI surfaces by qualified leads. Since Google Ads does not report AI Overview placements separately, track cost per qualified lead by campaign and query theme, before and after you widen matching.
- Don't reset targets because of a CTR change. A lower CTR on informational queries means fewer clicks for the same impressions. It doesn't mean your bids are wrong. Fix the query mix first.
Where ClickFortify fits
ClickFortify doesn't change how AI Overviews look or how often people click. What it does is deal with the clicks you pay for, whichever search surface they come from. ClickFortify for Google Search scores every click on dozens of behavioral, device and network signals in real time, adds fraudulent IPs to your Google Ads IP exclusions automatically, and covers Search, Shopping and Performance Max (the campaign types eligible for ads within AI Overviews) through Google's native API. Every block is logged with the signals behind it.
Keeping fake leads out of your primary conversions is still the job of your conversion setup, as described above. That is what stops Smart Bidding from learning from them.
