AI Overviews Displaced 25% of Your Ads. Here's a Budget Allocation Framework Built on the Data That Remains.

Paid CTR on informational queries with AI Overviews fell from 19.7% to 6.34% between June 2024 and September 2025, according to Seer Interactive's analysis of 3,119 queries across 42 organizations (reported by Search Engine Land). A 68% decline. Google offers no way to target AI Overview placements, no way to opt out, and no reporting breakout for ads served inside Overviews. They're lumped into "top ads."

You're losing click volume on a growing share of queries, you can't see where your ads actually ran, and the platform won't let you choose. The question isn't whether AI Overviews affect your paid search. It's which queries are still worth fighting for, which ones you should try to influence, and where paid search is the wrong channel entirely.

The Three-Bucket Framework: Fight, Influence, Generate Demand

A single bidding rule across all queries treats every click as the same opportunity. After AI Overviews, that assumption costs you money. Different queries now face structurally different economics on the SERP.

Fight is for bottom-of-funnel, high-intent queries where a click still feeds pipeline directly: product + modifier, "buy," "quote," brand terms with purchase intent. Winning absolute top position is worth paying for here, but only where your incremental CPA holds. Watch Absolute Top impression share and impression share lost to rank to see whether the Overview is pushing you below the fold.

Influence is for mid-funnel and informational queries where the AI Overview does the educating and the user often won't click. Your goal shifts from buying a visit to appearing inside the answer as a cited source or named brand. Seer Interactive's Q3 2025 data: paid CTR was 7.89% when the advertiser's brand was cited in the AI Overview, versus 4.14% when it wasn't. Brands cited in AI Overviews received 91% more paid clicks than brands not cited (Search Engine Land, citing Seer Interactive). That turns "Influence" into a coordination problem between paid and organic, not a bidding problem.

Generate Demand is for queries where zero-click has already won. Paid search alone can't rebuild the funnel. Budget moves upstream: YouTube, Demand Gen campaigns, first-party content, community presence. Capture happens later, in Fight, once the person searches with commercial intent. Measured on last-click ROAS, this always looks weak. Measured on brand search lift and qualified pipeline four to eight weeks downstream, it's where the pool gets filled.

Why Query-Level Segmentation Matters

Adthena's research, as reported by Optmyzr, suggests AI Overviews displace roughly 25% of ads. In healthcare specifically, ads appeared below AI Overviews 64.6% of the time (Search Engine Land, citing Adthena). Impact varies by industry, device, query type, and placement. A blanket policy like "pause all informational keywords" will be wrong for some segments. So will "bid up everywhere."

One wrinkle worth flagging: Seer's data showed paid CTR also declined on queries without AI Overviews. Other factors (SERP layout changes, user behavior shifts, competition) contribute. Attributing all performance changes to AI Overviews without query-level comparisons overstates the effect and leads to wrong budget moves.

How to Classify and Act

Pull your search terms and map intent: commercial, research, informational. Use Search Console, manual SERP checks, and whatever AI Overview monitoring you have. Sort each term into Fight, Influence, or Generate Demand. Revisit regularly; the SERP keeps changing underneath you.

On Fight terms, accept a position premium only where absolute top still beats your incremental CPA target. On Influence terms, align ad copy with the question the Overview answers and coordinate with SEO on content and citation strategy. Pull back paid spend where organic already earns visibility. On Generate Demand terms, redirect budget to YouTube, Demand Gen, or content. Measure through brand search volume changes and assisted pipeline, not last-click attribution.

The hypothesis (make it falsifiable): if you reclassify informational queries out of Fight and into Influence or Generate Demand, then blended CPA on remaining Fight terms will improve by 10–20% within four weeks, because you've stopped paying premium CPCs for clicks the SERP has already absorbed. Success = CPA improvement on Fight terms; guardrails = total qualified pipeline volume doesn't drop more than 10%; stop-loss = if pipeline drops 15% in week three, pause the reallocation and diagnose which terms were misclassified.

Google reports in-Overview ads as "top ads" with no breakout (Google Ads Help). Any claim that AI Overview ads specifically drove a KPI change is hard to prove with native reporting. Treat conclusions as directional. Build query-level segmentation outside Google's native tools, run holdouts where you can, and compare AI Overview-present versus absent queries manually until better reporting arrives.

The SERP already answered the user on a growing share of your keywords. The budget question is whether you're still paying for the click or investing in the conditions that make the next click worth more.