Ads in AI Overviews and AI Mode: The New Inventory

Ads in AI Overviews are sponsored results that Google places inside and alongside the AI-generated answers at the top of its results page. Ads in AI Mode extend the same idea into Google's conversational search surface, where the "query" may be a long, multi-turn question. Both are labelled Sponsored, and both are bought with the campaigns you already run — there is no AI-surface campaign type, no separate opt-in and no placement-level bid.

This is new inventory arriving under old plumbing, and the honest state of knowledge about it splits cleanly into confirmed and not.

How ads in AI Overviews work

Eligibility is the settled part. Ads from Search, Shopping and Performance Max campaigns can appear on AI surfaces, matched to the underlying query and the context of the AI answer. Because conversational queries are longer and stranger than typed keywords, the matching layer does the heavy lifting — which is why Google steers advertisers toward its broader-matching machinery (AI Max, broad match, Performance Max) as the practical route to this inventory. Google began placing ads in AI Overviews in the US in 2024 and has expanded surfaces and announced AI Mode placements since; rollout remains staged by country and surface.

What deserves a labelled ledger:

Confirmed	Still unknown from outside Google
Ads appear in and below AI Overviews, and in AI Mode, labelled Sponsored	Auction dynamics and pricing on AI surfaces
Eligibility via existing Search, Shopping and PMax campaigns	When placement-level reporting will exist
No separate opt-in, placement bidding or AI-surface campaign type	How click and conversion behaviour differs on these surfaces
Standard reporting does not break out AI-surface placements as of this writing	Rollout pace across countries and languages

The reporting row is the operational headache: your AI-surface impressions are folded into normal campaign metrics, so you cannot yet segment what this inventory does to your economics. Claims that pretend otherwise are guessing.

Managing a placement you cannot manage directly

You cannot manage this placement directly, so the work is managing around it — and that is mostly reading, which is what an agent is for.

First, watch the query mix. AI surfaces reward conversational intent, and if your traffic is shifting that way it shows up in your own search terms: longer queries, question forms, multi-clause asks. An agent pulling the search terms report on a schedule can quantify that drift instead of leaving it to impressions and anecdotes.

Second, watch the trend lines against the rollout. When CTR or CPC bends without any change in your account, surface changes at Google's end belong on the suspect list — an agent correlating your metrics with your own change history can at least rule you out, which is half the diagnosis. The wider 2026 context sits in what changed in Google Ads.

Third, apply a fraud filter. Because the inventory is new and opaque, it attracts confident selling — placement targeting that does not exist, AI-surface CPC benchmarks nobody outside Google could measure. An account-connected agent grounds those conversations in your data: what is confirmed, what is visible in this account, and what is somebody's pitch. The preparation that is actually in your control — eligibility, measurement, landing pages that answer conversational intent — is laid out in getting your ads into AI search.