Google Ads in AI Mode: What Early Testing Actually Shows

Google's AI Mode doesn't look like a search results page. There's no familiar blue-link grid, no obvious ad slot above the fold, and no clear visual cue that what you're reading was influenced by a paid placement. Yet ads are appearing inside it — and if you're watching click-through rate as your primary signal, you're likely drawing the wrong conclusions from the data you're already collecting.

The specific problem: AI Mode volume is arriving in accounts right now, it looks like underperformance in standard reports, and the natural response — tightening bids, narrowing match types — actively works against you. The rest of this post explains why, and what to do instead.

TL;DR

TL;DR — Ads in Google AI Mode

Google is serving Shopping and Search ads inside AI Mode and AI Overviews, not just on traditional results pages.
Click-through rates inside AI surfaces look lower than traditional search — but impression context is fundamentally different, so direct comparisons mislead.
Conversion signals matter more than click signals in AI Mode; accounts that chase CTR tend to bid themselves out of high-intent placements.
Smart Bidding campaigns (Performance Max, broad match with Target CPA/ROAS) are already eligible for AI Mode placement with no separate setup required.
The right immediate action is to audit your conversion tracking before touching bids — bad conversion data fed into AI surfaces compounds faster than it does in classic search.

What "Ads in AI Mode" Actually Means

AI Mode is Google's fully generative search experience — a conversation-style interface that synthesizes an answer rather than listing links. It is distinct from AI Overviews (the summary box that appears above traditional results on standard search), though both surfaces now carry ads.

The placement mechanics differ from what most advertisers are used to. In traditional search, your ad wins an auction for a fixed slot at the top or bottom of a page. In AI Mode, ads are woven into or displayed alongside a generative response. The user isn't looking at a list of ten results and choosing one — they're reading a synthesized answer, and the ad appears as a relevant commercial option within or adjacent to that context.

Google has confirmed that Shopping ads and Search ads are both eligible to appear in AI Mode, and that existing campaigns — particularly those using Smart Bidding — are already entering these auctions without any account-level toggle required. If you're running Performance Max or broad match campaigns with automated bidding, you are already in.

The opt-in you didn't notice

There is no "enable AI Mode placements" checkbox in Google Ads. Eligibility flows from your existing campaign structure and bidding strategy. If you're running Smart Bidding, you're already participating.

Why CTR Is the Wrong Metric to Watch

The instinct when a new surface launches is to pull a placement report, see lower-than-expected CTRs, and conclude the placement is weak. That instinct is wrong here, and it's worth being precise about why.

In traditional search, a low CTR on a high-intent query is a clear signal of copy or relevance failure. In AI Mode, the same CTR may reflect that the user got a substantial portion of their informational need met by the generative answer before they ever clicked anything. The impression itself is doing different work than a traditional impression does.

What testing and practitioner reporting show — corroborated by the YouTube breakdown at Ads in AI Mode for Google Ads — is that clicks from AI Mode surfaces tend to arrive from users who are further along in their decision process. They've already read a synthesized answer addressing their informational questions. When they click, it's a more deliberate action than clicking the first result on a traditional SERP.

Optimizing to suppress AI Mode impressions because the CTR looks weak is exactly backwards. You'd be reducing exposure at the placements where downstream intent is concentrated.

The Metric Shift: From Click Signals to Conversion Signals

If CTR misleads in AI Mode, conversion tracking becomes load-bearing in a way it wasn't before. Smart Bidding algorithms that power AI Mode eligibility train on your reported conversions. If your conversion events are misconfigured — firing on page load instead of form submission, double-counting, missing attribution for phone calls or offline actions — the model learns the wrong thing and mis-prices your bids in ways that are fast and hard to reverse.

The accounts that struggle most as Google's AI surfaces expand aren't struggling because AI Mode is broken. They're struggling because their measurement setup was already fragile and the AI amplifies fragility. A small tagging error that slightly overstated conversions in classic search could be absorbed. The same error fed into a generative-surface bidding model accelerates in the wrong direction faster.

Practical conversion audit checklist before you touch bids

Before adjusting any Smart Bidding targets in response to AI Mode data:

Verify your primary conversion action fires exactly once per qualified lead or purchase — load the confirmation page in a clean browser session and confirm one, not two, conversion pings appear in Tag Assistant.
Check for duplicate conversion imports — GA4 goals and Google Ads tags tracking the same event both imported as primary conversions is a common source of inflation.
If you have offline conversions (CRM imports, phone calls), confirm the upload cadence is current. Stale offline data starves the model.
Set non-primary conversions (PDF downloads, video views) to secondary status so they don't influence Smart Bidding targets.

Getting this right matters more now than it did six months ago.

What Campaign Types Actually Show Up in AI Mode

Not all campaigns are equally eligible. From what Google has disclosed and from practitioner reporting:

Shopping campaigns — both standard and those inside Performance Max — have the clearest presence in AI Mode, particularly for product queries. When a user asks a generative question implying commercial intent, Shopping-style ads surface as product options.

Search campaigns with broad match and Smart Bidding are eligible. Exact match campaigns with manual CPC have far more limited exposure because the bidding signal AI Mode relies on — real-time value estimation against a generative query — requires an automated bidding model to function properly.

Performance Max is the campaign type most structurally aligned with AI Mode. It was built to follow Google's signals across surfaces without the advertiser specifying placement by placement. AI Mode is one more surface from PMax's perspective.

Display-only and Video campaigns do not currently appear in AI Mode based on available practitioner sources and Google's own disclosures.

Manual CPC in AI Mode

If a large share of your Search budget runs on manual CPC with exact match, you may already have minimal AI Mode exposure — not because you're excluded, but because the placement favors campaigns that give Google's model room to operate.

Creative and Copy: What Actually Gets Pulled In

AI Mode doesn't display your ad the way a standard text ad appears. The generative interface can pull components from your ad — a headline, an image from your Shopping feed, a price — and present them in a format that fits the conversational layout. This makes asset quality more consequential, not less.

Headline specificity. Vague brand-forward headlines contribute little when the ad is rendered alongside a generative answer that already described the product category in detail. Specific headlines — exact model names, concrete value differentiators, numeric claims — give the rendering system something useful to surface.

Feed data quality for Shopping. Product title, price, and image are the primary assets pulled for Shopping placements in AI Mode. A product title stuffed with irrelevant keywords or a low-resolution image fails on a standard SERP too, but in a generative context where the ad competes with a paragraph of synthesized text, a weak image is especially costly.

Assets (formerly extensions). Sitelinks, structured snippets, and callouts remain eligible in AI Mode placements. Accounts with thin asset coverage leave presentation quality on the table.

How to Measure AI Mode Impact When Placement Reporting Is Limited

This is the operational gap that frustrates most advertisers right now: you can't pull a clean placement-level report for AI Mode the way you can for, say, Google Display Network placements. So how do you know if it's working?

The most reproducible approach is a controlled time-period comparison within your own account. Pick a campaign with stable conversion history, note the date AI Mode traffic began scaling in your vertical, and compare cost-per-conversion before and after — holding targeting and bids as constant as possible. If efficiency held or improved, AI Mode is likely contributing positively. If it degraded, that's your signal to audit conversion tracking first before assuming the placement is the problem.

A cleaner version of this is Google's own campaign experiments tool. You can run a base campaign against a draft variant that uses a tighter bidding target, measure conversion volume and CPA delta across both arms simultaneously, and get a statistically meaningful read without exposing your entire budget to the test condition. It's slow — you need several weeks of data — but it's the only method that controls for external seasonality.

What you should not do: draw conclusions from a single week of data during a surface transition. AI Mode traffic is still ramping, and short windows produce noisy reads.

What We're Watching Next

Two things remain genuinely unresolved.

First, attribution. When a user reads an AI Overview or an AI Mode response and converts later — same session, different session, or after a brand search — how does that conversion credit flow? Google's current attribution models were built around a link-click world. Generative-surface assists that don't produce a click won't show up in your attribution path at all, which means AI Mode may be doing meaningful assist work that surfaces as direct or organic in your reports. That makes it look more expensive than it actually is on a last-click view.

Second, query visibility. Traditional Search campaigns give you search term reports. AI Mode queries — because they're often conversational and long-form — may not surface in those reports with the same fidelity. That limits your ability to understand what's triggering your AI Mode impressions and apply negative keywords intelligently. Google has not fully resolved this yet. It's a real operational gap, and it's why building clean conversion infrastructure now — before the reporting catches up — is the right sequence.

FAQ

What are ads in AI Mode on Google? Ads in AI Mode are paid placements — primarily Shopping and Search ads — that appear within or alongside Google's fully generative AI Mode search experience. They are served through the same Google Ads auction system as traditional ads but are rendered in a conversational interface layout rather than a classic results page format.

Do I need to do anything special to show ads in Google's AI Mode? No separate setup is required. If you're running Smart Bidding campaigns — including Performance Max, or Search campaigns with Target CPA, Target ROAS, or Maximize Conversions — your campaigns are already eligible to appear in AI Mode placements. Manual CPC campaigns with exact match keywords have much more limited AI Mode exposure.

Why is my CTR lower in AI Mode compared to regular search? AI Mode impressions occur when a user is reading a generative answer. A significant portion of their informational need may be met by that answer before they click anything. Lower CTR in this context doesn't mean lower value — clicks that do occur tend to come from users who are further along in their decision process. Focus on conversion rate and cost-per-conversion rather than CTR.

Does Performance Max work in AI Mode? Yes. Performance Max is structurally well-suited to AI Mode because it uses Google's automated signals to place ads across surfaces without requiring the advertiser to manage placement-by-placement eligibility. AI Mode is treated as one more surface within PMax's scope.

How should I adjust my bidding for AI Mode? Don't adjust bids specifically for AI Mode in isolation — there is no placement-level bid modifier for it. Instead, ensure your conversion tracking is accurate (not inflated or undercounted) and let your Smart Bidding target do the work. The model will allocate toward AI Mode impressions when it estimates they'll convert at your target efficiency.

How can I measure whether AI Mode is helping or hurting my account? Compare cost-per-conversion across stable time periods before and after AI Mode traffic scaled in your vertical, holding targeting constant. For a cleaner read, use Google's campaign experiments tool to run a controlled A/B test over several weeks. Avoid drawing conclusions from single-week snapshots during a surface transition — the data is too noisy.

Are Google AI Overviews the same as AI Mode for ad purposes? They are distinct surfaces. AI Overviews is the summary box that appears on traditional search results pages. AI Mode is Google's fully generative, conversation-style search experience. Both carry ads, but AI Mode represents a more complete replacement of the traditional SERP experience. Your campaigns can appear in both; the same principles around Smart Bidding eligibility and conversion signal quality apply to each.

The most actionable thing you can do this week isn't to restructure your campaigns for AI Mode — it's to open Tag Assistant on your highest-spend conversion path and confirm the signal you're sending is actually clean. Everything else builds on that.