Google Tests Search Ads in AI Mode: Keyword Strategy

Introduction

For two years, the rise of AI Mode in Google Search has made advertisers and keyword researchers wonder whether traditional keyword-based campaigns still matter in the AI era. This week Google answered with a small but significant experiment: it started serving regular pure Search campaigns with exact and phrase match keywords inside AI Mode.

Confirmed by Google's Ads Liaison Ginny Marvin on September 4, 2026, the test means advertisers no longer need AI Max, Performance Max, or broad match with smart bidding to reach users inside Google's AI-powered search experience. For anyone building keyword strategy in 2026, that is a signal worth unpacking — because it tells us exactly where Google thinks keywords still fit in an AI-first search world.

The News: Google Tests Traditional Search Ads in AI Mode

The experiment was first spotted by AdSquire founder Anthony Higman, who said he confirmed the behavior across multiple campaigns and tests. Google then confirmed it was running what it calls a "small experiment that recently kicked off."

Here is what is actually happening:

Exact and phrase match keywords from traditional Search campaigns are now eligible to serve text ads inside AI Mode.
Eligibility is limited to queries where Google identifies "explicit and direct user intent." If the user's intent is clear and maps directly to your keywords, your existing ads may appear in AI Mode.
Campaigns do not need to be migrated to AI Max or Performance Max to get this exposure.
Google says AI Max and PMax for Search remain the products designed to capture complex, conversational intent and to access newer AI Mode ad formats like Highlighted Ads.

Before this experiment, the only ways to show up in AI Mode were through AI Max, PMax, or broad match combined with smart bidding. Restrictive match types effectively had no path into AI Mode. That is why the PPC and SEO communities reacted so strongly to the news — it is the first time classic keyword targeting has been invited back into Google's AI search surface.

Why This Matters for Keyword Researchers

If you work with keywords — whether for SEO, paid search, or social content — this experiment matters for three reasons.

1. Keywords are not dead; precise intent is the new currency

For months, the prevailing narrative has been that conversational AI search would make keyword targeting obsolete. Google's AI Mode experiment says the opposite: exact and phrase match keywords are valuable precisely because they encode clear intent. Google is testing whether tightly targeted Search campaigns can participate in AI Mode when a user's query maps directly to an advertiser's keywords.

The takeaway: keywords built around direct, commercial intent — "buy," "best price," "near me," specific product names, comparison phrases — are the ones most likely to surface in AI Mode. Fuzzy, generic, low-intent keywords are exactly the ones Google's AI will handle with an answer, not an ad.

2. Match type strategy still matters

Because the test is limited to exact and phrase match, advertisers using restrictive match types now have a path into AI Mode without broadening their campaigns. That means:

Exact match campaigns targeting high-intent commercial keywords could start seeing AI Mode impressions.
Phrase match gets a similar opportunity when intent is explicit.
Broad match + smart bidding remains the route for discovery-style, conversational queries — but it no longer has a monopoly on AI Mode visibility.

For keyword research, this re-opens the value of finding the exact phrases real users type when they are ready to act — and building separate, tightly themed campaigns around them.

3. AI Mode is becoming a monetized surface

Google is clearly investing in AI Mode as an ad surface. The same week as the Search ads experiment, Google was reported to be testing two shopping ad placements inside AI Mode results, and earlier in September it added Gemini 3.8 Flash to AI Mode's model menu for Pro and Ultra subscribers. When Google starts testing ads inside a product, you can assume the product is strategic.

The direction is consistent: AI Mode is shifting from a pure organic answer engine to a hybrid results page where AI-generated answers coexist with text ads, shopping placements, and eventually richer ad formats. Keyword researchers who track where the "ads" show up in AI answers will have a head start on everyone measuring only classic SERPs.

The Reverse Crocodile Effect: What the Data Shows

The same week, new data from Mike Ryan, head of ecommerce at Smarter Ecommerce, pointed to another AI Mode shift that keyword researchers should understand: what Ryan calls the "reverse crocodile effect" for Shopping ads.

Here is the context. The "crocodile effect" describes what AI Overviews did to publishers: impressions went up, but clicks went down — the crocodile's mouth is open wide (lots of visibility) but the teeth close on the click. For Shopping ads, the pattern appears to be the opposite:

Click-through rates are rising — Optmyzr data showed Shopping CTR up about 17% year over year, close to the roughly 20% increase Ryan was seeing in Smarter Ecommerce's Market Observer benchmark data.
Impressions are falling — across thousands of Shopping and Performance Max campaigns spanning hundreds of accounts, clicks were often flat or slightly down while impressions dropped more noticeably.
Net result: higher CTR, but fewer ads shown.

Ryan's hypothesis: Google may be more likely to serve AI Overviews for queries with a lower predicted likelihood of generating an ad click, while preserving Shopping ad placements for searches with stronger click propensity. He calls this "experience switching" — Google serves different experiences depending on the query, and anecdotally, Shopping ads and AI Overviews rarely appear together on the same results page. It is one or the other.

What does this mean for keyword strategy? The queries that still show traditional ad formats are the high-commercial-intent ones. The more Google funnels lower-intent queries into AI Overviews, the more the remaining ad inventory concentrates on queries that signal purchase readiness. That reinforces point one: intent-dense keywords are where the remaining traditional traffic lives.

How to Adapt Your Keyword Strategy

Audit for explicit-intent keywords

Go through your keyword lists and separate them into two tiers:

Tier 1 — explicit, direct intent: product + modifier ("wireless earbuds noise cancelling"), brand + product, "best X for Y," "X vs Y," "buy X," "X price," "X near me," "X delivery." These are the keywords most likely to trigger traditional ads in AI Mode under the current experiment.
Tier 2 — conversational and research intent: questions, "what is," "how to," comparisons without purchase modifiers. Google's AI will increasingly answer these with an AI Overview — plan for brand visibility in the answer, not for a click.

Build exact match campaigns around your money keywords

If the experiment expands beyond its current limited state, exact match campaigns will be the first to benefit. Make sure your highest-value commercial keywords exist as exact match keywords in dedicated, tightly themed ad groups — not buried inside broad match campaigns where the algorithm decides what your keywords mean.

Track AI Mode visibility separately

AI Mode has its own model menu (Gemini 3.8 Flash arrived September 2, 2026 for Pro and Ultra users), its own citation behavior, and now its own ad placements. If you are not already measuring how often your brand appears in AI Mode answers — and which queries trigger AI Mode versus classic results — start now. A keyword can rank on the classic SERP and be invisible in AI Mode, or appear in AI Mode for queries you never optimized for.

Watch Shopping impression trends in your accounts

If you run Shopping or Performance Max campaigns, do not celebrate rising CTR alone. The "reverse crocodile effect" data suggests higher CTR can simply mean Google is showing your ads less often, on fewer, higher-intent queries. Track impressions, clicks, and absolute conversions alongside CTR. A campaign with declining impressions may be losing ground to AI Overviews even as its CTR looks healthier.

Keep harvesting question and long-tail keywords

Even though Google is concentrating traditional ads on explicit-intent queries, question keywords and long-tail phrases still matter enormously — just for a different goal. They are the keywords that earn you a mention inside AI answers, and AI citations are increasingly the top-of-funnel traffic source. Product pages capture about 24% of AI citations in some analyses, with forums like Reddit and YouTube trailing — so structure content around the questions your customers actually ask.

Refresh keyword data monthly, not quarterly

The AI Mode model menu changed three times in six weeks (Gemini 3 Flash, then 3.5 Flash at I/O, then 3.7 Flash in August, then 3.8 Flash in September). Every model change can alter which results AI Mode shows and how it interprets queries. Keyword landscapes that used to shift quarterly are now shifting monthly. Build a habit of refreshing your keyword research at least once a month, and re-check your AI Mode visibility after every Google model or product announcement.

Key Takeaways

Google confirmed a small experiment (September 4, 2026) serving traditional Search ads with exact and phrase match keywords inside AI Mode, limited to explicit and direct user intent.
Advertisers no longer need AI Max, PMax, or broad match + smart bidding for a path into AI Mode — restrictive match types are being tested.
The "reverse crocodile effect" data suggests Shopping ad impressions are falling while CTR rises, as Google concentrates ads on higher-intent queries and answers the rest with AI Overviews.
Keywords are not dead — explicit-intent commercial keywords are becoming more valuable as the AI era concentrates monetizable clicks on them.
Audit your lists into explicit-intent and conversational tiers, keep exact match campaigns around money keywords, and track AI Mode visibility separately from classic SERPs.
Refresh keyword research monthly — AI Mode's model and feature set changed three times in six weeks.

FAQ

Q: Can my existing Search ads really show in AI Mode?

A: In the current limited test, yes — if you use exact or phrase match keywords and Google identifies explicit, direct user intent in the query. Google confirmed the experiment on September 4, 2026. It is a small test for now, and Google has not said whether it will become permanent or how broadly it is running.

Q: Do I need Performance Max or AI Max for AI Mode?

A: Not for this test. Exact and phrase match campaigns are being served in AI Mode without AI Max or PMax. However, Google still positions AI Max and PMax for Search as the products for complex, conversational intent and for newer AI Mode ad formats like Highlighted Ads.

Q: Does this mean AI Overviews are bad for my Shopping ads?

A: Not necessarily bad — but the data shows Shopping impression volumes are falling while CTR rises, a pattern researchers call the "reverse crocodile effect." It may mean Google shows Shopping ads less often and only on higher-intent queries. Judge campaign health on impressions and absolute conversions, not CTR alone.

Q: Is AI Mode the same as AI Overviews?

A: They are related but different surfaces. AI Overviews appear within classic Google results, while AI Mode is Google's dedicated AI-first search experience with its own model menu and answer formats. Both are expanding and both now interact with ads differently.

Q: How should keyword research change for AI Mode?

A: Split your strategy: target explicit-intent commercial keywords for ad and classic visibility, and conversational/question keywords for AI answer citations. Refresh monthly — AI Mode's underlying models and features changed three times in six weeks during mid-2026.

Conclusion

Google's decision to test traditional Search ads in AI Mode is one of the clearest signals yet about the future of search monetization: AI Mode is becoming a real ad surface, and keywords with explicit intent are the bridge between classic search and AI search.

For keyword researchers, the message is encouraging. The death of the keyword has been exaggerated — but the type of keyword that matters is changing. High-intent, precise, commercially relevant phrases are gaining a second life inside AI Mode, while fuzzy research queries migrate to AI-generated answers where brand visibility, not clicks, is the goal.

Stay ahead of the shift by auditing your keyword lists for explicit intent, keeping exact match campaigns around your money keywords, tracking AI Mode visibility separately, and refreshing your research monthly. The sites and advertisers that treat AI Mode as a distinct, measurable surface — with its own keywords, its own ads, and its own citation behavior — will be the ones still growing when the next model update lands.

Ready to find the high-intent keywords your competitors are missing? Generate fresh, intent-focused keyword lists in seconds at SocialKeywordGenerator.com and build the exact match campaigns that will perform in both classic Search and AI Mode.