AI Max for Search
Not Performance Max, and not something you necessarily chose to turn on. Here's exactly what it is, what it needs to work well, and the honest range of results to expect — not Google's best-case study.
Rohan Alexander · 11 min read · Updated July 2026
Quick Answer
What AI Max for Search Actually Is
AI Max is not a new campaign type — it's a set of AI-powered features layered on top of your existing Search campaigns. It expands keyword matching (finding query variations beyond your literal keyword list), can generate ad copy variations automatically, and in some configurations sends traffic to landing pages on your site that you didn't explicitly specify in your ad group, based on Google's own model of relevance. Crucially, it preserves what standard Search campaigns already give you: full search term reports with source attribution, keyword-level bid control, and the ability to add negative keywords at the campaign and ad group level — none of which Performance Max offers.
AI Max vs Performance Max
| AI Max for Search | Performance Max | |
|---|---|---|
| Scope | Search inventory only | Search, Shopping, Display, YouTube, Gmail, Discover, Maps |
| Search term visibility | Full search term reports retained | Very limited query-level transparency |
| Negative keywords | Campaign and ad group level, same as standard Search | Campaign-level only (expanded from 100 to 10,000 slots in March 2025) |
| Structure | Layers onto existing Search campaign structure | Standalone campaign type, asset-group based |
| Best fit | Accounts that have optimized standard Search and want incremental volume | Businesses wanting reach across all Google properties with strong creative assets |
Running both together is common — the key is preventing overlap, since both can compete for the same queries and inflate blended cost. Segmenting by intent (AI Max for high-intent bottom-funnel search, PMax for broader awareness and discovery) and adding campaign-level negatives to keep AI Max from competing with PMax for identical terms is the generally recommended approach.
Account Prerequisites Before Testing
| Prerequisite | Why it matters |
|---|---|
| Enhanced Conversions enabled | AI Max's expanded matching needs accurate, complete conversion signal to optimize against |
| Target CPA or Target ROAS bidding, not Maximize Conversions | Value-based bidding goals give the expansion a clearer target than raw conversion volume |
| A meaningful monthly conversion volume on the campaign | Thin conversion data makes it hard to tell signal from noise once matching broadens |
| Daily budget comfortably above your target CPA | A tight budget relative to CPA limits how much the algorithm can actually test and learn from |
| A comprehensive existing negative keyword list | Expanded matching needs a strong existing negative list as a first line of defense |
Testing without these in place doesn't just risk wasted spend — it makes the test itself hard to interpret, since you won't be able to tell whether a disappointing result reflects AI Max specifically or an account that wasn't ready for any bidding change.
Realistic Results, Not the Best-Case Study
Google's own case studies cite large headline improvements, but independent testing across many advertisers tells a more modest story — reported average conversion improvements in the low double digits, with the strongest lifts concentrated in accounts that were previously running heavily on exact and phrase match (where AI Max fills a real coverage gap). A modest, genuine improvement over your existing baseline is a realistic win; comparing your own results to Google's most favorable case study sets an unrealistic bar and can make a perfectly reasonable outcome feel like a failure.
Checking If It's Already Active on Your Account
Google has enabled some AI Max-related features by default on new Search campaigns, and has auto-enrolled some existing campaigns into certain features through interface updates — meaning it's entirely possible AI Max is already influencing an account without anyone having explicitly opted in. Check campaign settings directly under "AI-powered features" and broad match/keyword expansion settings rather than assuming the current setup reflects a deliberate choice.
How to Test It Properly
- Pick your strongest existing Search campaign — established conversion volume, stable performance — not a struggling one you're hoping AI Max will rescue.
- Use Google's experiment feature to run a controlled comparison against your existing setup, rather than switching the whole campaign over blind.
- Run the test for at least 4 weeks to allow the expanded matching to actually learn, checking search term reports daily in the first week specifically to catch irrelevant queries early.
- Only roll out further once the test shows a genuine, positive result — not a hunch or a single good day.
Negative Keywords Still Matter — More Than Ever
There's no toggle to fully disable AI Max's keyword expansion once features are active — the only real defense against irrelevant matches is a robust, actively maintained negative keyword list, reviewed regularly rather than set once. See Negative Keywords: The Complete Strategy Guide for how to build and maintain this systematically, since AI Max's expansion rate can outpace a negative list that's only updated occasionally.
Text Guidelines: Keeping Control Over Ad Copy
Text guidelines let an advertiser set explicit guardrails on AI-generated ad text — tone, phrases to avoid, claims that must not be made — addressing one of the most common early objections to AI Max: losing control over what the ad actually says. This feature reached global availability for all advertisers in early 2026 and is worth configuring before enabling AI-generated asset creation, not after an off-brand variation has already run.
Variations by Account Maturity
| Account maturity | AI Max approach |
|---|---|
| New or low-conversion-volume account | Wait until conversion volume and tracking are solid before testing |
| Established, exact/phrase-match heavy | Strong candidate — this is where the biggest reported lifts concentrate |
| Already running broad match aggressively | Smaller incremental gain likely; negative keyword discipline matters even more |
Case Study
A B2B services account running exclusively on exact and phrase match keywords, with solid Enhanced Conversions tracking and Target CPA bidding, tested AI Max on its top-performing campaign via a Google Ads experiment for five weeks. The test showed a genuine, if modest, improvement in conversion volume at a similar CPA to the control — consistent with the pattern of AI Max helping most where a campaign has real untapped query coverage. A newer campaign, still building conversion history, showed inconclusive results in the same test window, reinforcing that AI Max needs a real prerequisite state, not just activation, to show its benefit.
Decision Matrix
| Situation | Priority |
|---|---|
| Strong existing Search campaign on exact/phrase match | Good candidate for a controlled AI Max experiment |
| New account, thin conversion history | Build conversion volume and tracking accuracy first |
| Already stretched thin on negative keyword maintenance | Strengthen that process before adding a broader matching layer on top |
Common Mistakes
- Enabling AI Max across every campaign simultaneously instead of testing on one first.
- Comparing your own results against Google's best-case study rather than your own baseline.
- Assuming AI Max isn't active because it was never explicitly turned on.
- Letting the negative keyword list go stale once expanded matching is active.
Troubleshooting
Irrelevant queries appearing after enabling AI Max: check search term reports daily for the first two weeks and add negatives immediately — don't wait for a monthly review cycle.
Unsure if AI Max is already affecting your account: check campaign settings under "AI-powered features" directly rather than assuming based on when the account was created.
Checklist
☐ Enhanced Conversions enabled and verified accurate
☐ Target CPA or Target ROAS bidding in place, not Maximize Conversions
☐ Meaningful monthly conversion volume on the test campaign
☐ Negative keyword list comprehensive before testing
☐ Test run via experiment feature for at least 4 weeks
☐ Text guidelines configured before enabling AI-generated ad copy
AI Prompts to Speed This Up
- "Review this search term report [paste] from an AI Max test and flag queries that look like they warrant a negative keyword."
FAQ
What is AI Max for Search?
An AI expansion layer inside standard Search campaigns that broadens keyword matching while keeping your structure, search term reports, and negatives intact.
How is it different from Performance Max?
AI Max stays within Search with full query transparency; Performance Max spans all Google properties with far less visibility.
What prerequisites should be in place before testing?
Enhanced Conversions, Target CPA/ROAS bidding, meaningful conversion volume, adequate budget, and a strong negative keyword list.
Is AI Max already active on my account without me turning it on?
Possibly — check campaign settings directly rather than assume.
Testing AI Max properly means daily search term monitoring for weeks. Most accounts don't have that time.
Zephra runs controlled AI Max experiments, reviews search term reports daily, and adds negative keywords the moment query quality drops — not on a monthly review cycle.
Start Free Audit →Sources & Further Reading
- Google Ads Help — About Negative Keywords — Official documentation on negative keyword behavior across campaign types including AI Max and Performance Max.
- WordStream — 2026 Google Ads Benchmarks Report — Current cross-industry CPC, CTR, and conversion rate benchmarks.
Figures referenced in this guide (including performance ranges and thresholds) are drawn from independent 2026 advertiser reporting, not Google's own published guarantees — confirm current settings and eligibility directly in Google Ads before making budget decisions.