Log in / Create free account🌐 ES☀️

Advanced training · Module 3 — AI Max — the black box of automatic Search expansion

AI Max for Search: the promise, the components and the part you can't see

⏱️ 9 min read · 🔎 Search 📐 Measurement · updated on 2026-09-02

🎧 Listen to the lesson · ≈ 4 min🔒 Subscribers only

In short: AI Max adds dynamic broad matching, keywordless traffic and generated text and URLs to your Search campaigns, and often switches itself on by default and in bulk. Its problem isn't performance — which varies — but traceability: in real accounts up to a third of the spend shows up as "no breakdown". The opacity is constant, so you have to measure from outside.

AI Max is the latest layer of automation Google lays over Search campaigns, and the one that most changes the relationship between what you configure and what you pay for. This module develops the technical study «The black box of AI Max» from the Library: two real accounts, opposite results, and a method for knowing which of the two cases you are in. This first lesson introduces the tool as Google describes it — and the part it doesn't describe.

What is AI Max in Google Ads?

AI Max is a set of features you switch on in an existing Search campaign (it is not a new campaign type). Three components:

Component What it does
Dynamic broad matching Your keywords (in any match type) can trigger related searches beyond their match type, using the criteria of modern broad match
Keywordless traffic The campaign enters searches that hang off none of your keywords, based on your landing page, your ads and your history
Asset customization Headlines and descriptions generated by AI for each search (what used to be "automatically created assets") + final URL expansion (it sends the click to whichever page of your site it thinks is best)

The promise: incremental volume — queries your keywords did not cover, with ads adapted to each one. Google reports average uplifts in conversions when you switch it on, especially in campaigns with exact and phrase match.

How do you switch AI Max on and off?

In the Search campaign's settings: an AI Max toggle, with its components switchable separately (matching, text customization, URL expansion) and with controls: excluded brands, URL exclusions, negatives, and in some cases locations of interest. Important: in many accounts it has turned up already switched on, or has been rolled out in bulk across dozens of campaigns on the same day — the study documents it: 14 of the 21 campaigns in one account started generating AI Max traffic on the same day, 12 May 2026. If you don't remember switching it on, go and look.

What data does Google show about AI Max?

The AI Max reports show search terms with their source (classic keyword, dynamic broad, keywordless), the generated headlines and the URLs it sent the click to, and performance by campaign. It is more detail than PMax gives on Search; and, as we shall see, less than it looks.

What can't you see about AI Max?

The problem with AI Max is not that it performs badly: sometimes it performs very well. The problem, and it is a measurable one, is traceability: Google does not clearly separate AI Max's performance from your keywords', and the reports that ought to let you audit it are trimmed by privacy thresholds that are different from one another. The consequence, to preview it: in the study, one real estate account could reconcile only 2 of the 200 highest-spending combinations (1%) with the search terms report; in a training account where AI Max performed exactly as well as the keywords, a third of the spend still ended up in a "no breakdown" bucket that Google attributes to nothing.

The thesis of this module: the opacity is constant; the result is variable. From the default reports, the good account and the bad one are indistinguishable. Only independent measurement separates the case where you need to act from the case where you don't.

What this module covers about AI Max

  1. The mechanism of the opacity (three layers stacked on top of each other).
  2. How to measure what you can't see (separating by source, before/after with a control, reconciling).
  3. The two real cases, with the numbers.
  4. What the advertiser can do.
  5. How AI Max Analyzer (free in the panel) automates it.

💡 Ninja trick: the first action, before any analysis, is knowing when AI Max was switched on in each campaign. The search terms report with the source column and the change history tell you; AI Max Analyzer detects it on its own (the first date with AI Max traffic per campaign) and sets the baseline there: without that date, no "it's going well" or "it's going badly" has anything to be compared against.

What you should remember

📎 Sources and further reading

⚠️ Free training with no support. Ninja Scripts support channels (email and Telegram) are only for the use of the scripts, not for Google Ads questions or questions about this training.

Pick up here

After →Opacity by design: the three layers that switch off the lightAI Max — the black box of automatic Search expansionRelacionadaThe columns that matter: recommended sets by levelInside Google Ads — reports, columns and segmentsRelacionadaPredefined reports, the report editor and auction insightsInside Google Ads — reports, columns and segmentsRelacionadaThe 10 metrics you have to master before spending a single euroHow advertising on Google worksRelacionadaCase: long-cycle B2B lead generation — few leads, high value and a funnel that lasts monthsAccount architectures for scale — real cases

Ver el temario completo

🎓
You're reading the Ninja Academy in the open

Create your free account (no card) to save your progress, take the graded quizzes, earn the Basic level certificate and unlock part of the Intermediate and Advanced training.

Create my free account →
🥷

Subscriber feature

This option is part of the Ninja Scripts Suite subscription.

See the subscription →