Three signals dropped on the Google Ads ecosystem this week. They look like routine industry updates if you skim them. They point in the same direction if you read them together.
Each comes from a different vantage point. The first is operator data from a third-party platform with millions of active accounts. The second is product strategy from the senior person running Google Search. The third is engineering guidance from inside the Google Ads team itself. The convergence between them tells you what is actually happening underneath the auction.
Signal One: CTR Is Rising. Conversions Are Flat.
Optmyzr published a report this week analyzing performance trends across its account base. The headline finding was clean and uncomfortable. Click-through rates on Google Ads campaigns are trending up. Conversion rates are not moving with them.
The implication is structural. The auction is increasingly delivering clicks that match query intent at a surface level but do not match purchase intent at a depth level. The AI-driven ad placement layer is optimizing for engagement signals (CTR, time on landing page, scroll depth) that do not perfectly correlate with the conversion outcomes advertisers care about. The gap is small enough to miss in a single account audit and large enough to materially distort budget allocation when it shows up across thousands of accounts.
The operational consequence is that ROAS-only optimization gets less reliable every quarter. The advertisers I have talked to who are managing this well are layering customer-quality signals into their conversion definitions, not just transactional volume. They are also rebuilding their bidding strategies around lifetime value cohorts rather than immediate-conversion ROAS, which gives the AI more useful signal to optimize against.
Source: Optmyzr report on Search Engine Journal.
Signal Two: Different Queries, Different Surfaces
Liz Reid, Google's Head of Search, gave a separate Search Engine Journal interview this week explaining that AI Mode and classic SERPs serve different intent classes. Browsy, exploratory queries continue to perform better with full classic results. Complex, follow-up, multi-step queries perform better with AI synthesis.
The advertising implication is that paid placement now lives in a fragmented surface environment. Your ads run differently inside an AI-mediated answer than they do inside a classic blue-link SERP. The visibility logic, the click logic, and the conversion logic all shift depending on which surface the user landed on. Most advertisers are still measuring as if there was one surface.
The audit move I would push is clear. Segment your campaign reporting by surface (classic SERP vs AI Mode vs Search Live) wherever the platform supports it. The performance differences across surfaces are usually large enough to redirect budget meaningfully. Treating them as a single channel is the kind of operational laziness that keeps ROAS reports looking healthy while the underlying mix degrades.
Source: Search Engine Journal interview with Liz Reid.
Signal Three: Google Wants You to Use AI Differently
The third signal came from inside the Google Ads engineering team. A senior engineer made the case publicly that advertisers should use AI to enhance the value of their offering, not to scale output volume. Quoted by Search Engine Journal, the framing was that the ad ecosystem benefits when AI is used for differentiation rather than for production scaling.
This is the most direct guidance the platform has given on how it intends to weight quality over volume in the coming auction logic. The message between the lines is that the spam-style scaled-content tactics that some advertisers have been running with AI tools are going to get penalized, not rewarded. The auction is being built to favor advertisers who use AI to make their offer more relevant, more accurate, and more useful.
Read together with the other two signals, the picture clarifies. CTR is rising because AI ad placement is getting better at finding engagement. Conversions are flat because the engagement is not perfectly tied to purchase intent. Different surfaces produce different conversion behavior and most accounts are not segmenting correctly. And Google itself is signaling that the way out is using AI for differentiation rather than scaled production.
Source: Search Engine Journal coverage of Google engineer guidance.
Three signals. Three vantage points. One direction.
The advertisers who treat them as a coherent message will reallocate inside Q2. The ones who treat them as separate news items will be wondering by Q3 why their ROAS reports stopped predicting actual revenue.
FAQ
Why are Google Ads click-through rates rising while conversion rates stay flat?
The AI-driven ad placement layer is increasingly optimizing for engagement signals (CTR, time on landing page, scroll depth) that correlate with surface-level intent matches but not perfectly with purchase intent. The auction delivers more clicks but the marginal click is less likely to convert. The fix is layering customer-quality signals into conversion definitions and bidding around lifetime value cohorts rather than immediate-conversion ROAS.
How should advertisers report performance across AI Mode, classic SERPs, and Search Live?
Segment campaign reporting by surface wherever the platform supports it. The visibility, click, and conversion logic differ significantly across the three surfaces, and treating them as a single channel masks performance variation that should drive budget reallocation. Most accounts are still reporting as if there was one search surface, which produces ROAS reports that look healthy while the underlying mix degrades.
What does Google's recent engineering guidance mean for AI use in ad production?
The auction logic is being weighted to favor advertisers who use AI to make their offer more relevant, accurate, and useful, not advertisers who use AI to scale generic output volume. The spam-style scaled-content tactics some advertisers have been running with AI tools are likely to get penalized rather than rewarded. The strategic move is using AI for differentiation work that improves quality, not for output volume that floods the auction.