Google Ads AI Max for Search: What It Is and How to Use It Safely

Google Ads AI Max (officially "AI Max for Search campaigns") is a bundle of AI features you switch on inside a regular Search campaign. It does three things: it matches your ads to searches you never added as keywords (using your landing pages, ads, and existing keywords as the signal), it generates headlines and descriptions on the fly, and it can send the click to a different page on your site than the one you set. It is not a new campaign type and not a new match type in the strict sense, although it shows up as "AI Max" in your search terms report. Google announced it in May 2025, took it out of beta in 2026, made it the default for new Search campaigns, and is now migrating legacy broad match and Dynamic Search Ads setups into it. If you run Search campaigns, you'll be dealing with it whether you opt in or not, so this guide covers what it is, when it helps, when it hurts, how to set it up with guardrails, and how to audit it in minutes with AI.

Key takeaways

• AI Max is three toggles, not one: search term matching, text customization, and final URL expansion. You can turn each one on or off separately, and search term matching can be disabled per ad group.
• It needs a conversion-based Smart Bidding strategy (Maximize conversions, Maximize conversion value, with or without a target). Without one, search term matching won't work.
• Google's own numbers: a typical lift of 14% in conversions or conversion value at a similar CPA/ROAS, up to 27% for campaigns built mainly on exact and phrase match. Independent tests are mixed, especially for small budgets.
• The controls matter more than the toggle: brand inclusions/exclusions, URL exclusions, locations of interest, text guidelines, and negative keywords are what keep it from spending on junk.
• Deadline reality: Google stopped letting you create new campaign-level broad match and legacy automatically created assets in August 2026 and migrated them to AI Max in September 2026. Dynamic Search Ads follow in 2027.

What Google Ads AI Max actually is (and isn't)

Think of AI Max as a layer on top of a Search campaign. Your keywords, ad groups, ads, and bidding stay where they are. AI Max adds automation in three places.

1. Search term matching

This is the big one. When it's on, two things happen:

• Your exact and phrase keywords behave like broad match. A campaign built entirely on `[running shoes for flat feet]` can now show for "best sneakers overpronation" if Google decides the intent matches.
• Keywordless matching kicks in. Google reads your landing pages, your existing ad copy, and your keyword list, then enters auctions for searches that have no keyword behind them at all. This is the same technology that powered Dynamic Search Ads, now inside a normal campaign.

Search term matching is on by default when you enable AI Max. You can switch it off at the ad group level, which is useful if you want brand ad groups or high-value exact-match groups to stay tightly controlled.

2. Text customization

This is the renamed version of the old "automatically created assets" setting. Google generates new headlines and descriptions from your landing page, your other ads, and your keywords, then mixes them with your own assets in responsive search ads. You can review the generated assets, remove ones you don't like, and set text guidelines (more on that below).

3. Final URL expansion

Google can send the click to whatever page on your site it considers the best match for the query instead of the final URL you set. For a store with 2,000 products that's often a good thing: someone searching a specific model lands on that product page rather than your category page. For a lead-gen site with one conversion-optimized landing page, it's often a bad thing. You can exclude URLs or restrict expansion to a list of pages.

AI Max vs broad match vs Performance Max vs DSA

People mix these up constantly, so here is the short version:

• Broad match is a keyword match type. It still needs a seed keyword. Keyword-level broad match still exists; the campaign-level "upgrade everything to broad match" setting is what got folded into AI Max.
• Dynamic Search Ads (DSA) were a separate ad group type that used your website to pick queries and write headlines. AI Max's keywordless matching plus text customization replaces that, and Google has said DSAs will be migrated to AI Max in 2027.
• Performance Max is a separate campaign type that runs across Search, Shopping, YouTube, Display, Discover, Gmail, and Maps from one asset group. AI Max stays in Search only, keeps your campaign structure, and gives you keyword-level and search-term-level reporting that Performance Max doesn't.
• AI Max sits inside your Search campaign. You keep negative keywords, ad groups, and per-keyword data, but hand Google the decision on which queries to enter and what copy to show.

Microsoft Advertising launched its own feature with the same name and the same three components in August 2026, replacing its Predictive Matching setting. If you run Microsoft Ads too, the same playbook applies there.

What Google says it delivers vs what advertisers see

Google's headline claim at launch was a typical 14% more conversions or conversion value at a similar CPA or ROAS, rising to around 27% for campaigns running mostly exact and phrase match. That second number makes sense: a tightly restricted campaign has the most unused reach to unlock. By mid-2026 Google said more than 500,000 advertisers had adopted AI Max and lumped AI Max and Performance Max together at an average 15% lift in conversions or value at similar ROAS.

Independent results are less tidy:

• Several agencies published early tests showing a large share of AI Max impressions produced zero conversions, with incremental volume coming at a higher CPA than the keyword-driven traffic in the same campaign.
• Reports of expansion into Search Partner Network traffic and loosely related queries were common enough that the "AI Max" match type label in the search terms report (added July 2025) became the first thing experienced buyers check.
• In an October 2025 webinar Google itself framed the floor: around $50/day budget, a one- to two-week learning period, and consolidating ad groups that get fewer than 30 conversions a month. Google also says AI Max won't be effective in a campaign that is limited by budget, and shows an alert when it is.

The honest read: AI Max is a reach tool. If your campaign is capped by budget and already converting well, more reach means spending the same money on worse queries. If your campaign has budget headroom and a solid conversion signal, it usually finds incremental volume. Where it lands for you depends almost entirely on the controls and the conversion data you feed it.

Before you turn on AI Max: four prerequisites

Skipping these is how AI Max gets a bad reputation in an account.

• Accurate conversion tracking with values. Smart Bidding optimizes toward whatever you call a conversion. If "newsletter signup" and "purchase" both count as primary conversions, AI Max will happily find cheap signups. For ecommerce, send purchase value (ideally profit, see below) and bid on Maximize conversion value with a target ROAS.
• A conversion-based bidding strategy. Manual CPC and Maximize clicks don't qualify for search term matching.
• Budget headroom. If the campaign shows "Limited by budget," fix that first or AI Max simply redistributes your spend to broader queries.
• A clean negative keyword list. Everything you'd never want to pay for (competitors you don't want to bid on, "free," "jobs," "manual," "repair") goes in before you flip the switch, not after.

How to set up Google Ads AI Max step by step (manual)

Do it by hand once. You'll understand exactly which lever does what, and that makes the AI shortcuts later much safer.

Step 1: Pick the right campaign

Start with one Search campaign that has 30+ conversions a month, a Smart Bidding strategy, and is not limited by budget. Don't start with your brand campaign.

Step 2: Enable AI Max

• Open the campaign and go to Settings.
• Find the AI Max section (in newer accounts it sits alongside asset optimization settings; new campaigns created since 2026 have it on by default).
• Toggle AI Max on. You'll see the three sub-features: search term matching, text customization, and final URL expansion. All three default to on.

Step 3: Decide per feature

• Search term matching: leave on for the campaign, then go into any ad group you want to protect (brand, top exact-match products) and turn it off there.
• Text customization: leave on, but plan to review generated assets weekly for the first month.
• Final URL expansion: on for ecommerce catalogs with good product pages. Off, or restricted with URL inclusions, for lead-gen with a single landing page.

Step 4: Add the controls

• Brand inclusions: if this campaign should only show for queries mentioning your brand (or a partner brand), add them. Brand inclusions make the campaign brand-only.
• Brand exclusions: add competitor brands you never want to pay for. For a Shopify store, that's typically the big marketplaces plus direct competitors.
• URL exclusions: your blog, careers page, policies, out-of-stock collections, and any page without a buy button.
• Locations of interest (ad group level, exclusive to AI Max): lets you target people searching *about* a place rather than people *in* it. Useful for travel, local services, and shipping-restricted products. Skip it if location doesn't matter to you.
• Text guidelines: up to 25 excluded terms and 40 messaging restrictions per campaign. Add words legal won't allow ("guaranteed," "cure," "cheapest"), trademark terms, and claims you can't back up.

Step 5: Set the measurement baseline

Before you save, note the campaign's last 30 days: cost, conversions, conversion value, CPA/ROAS, and impression share. Better still, run it as a campaign experiment (Experiments > Custom experiment, 50/50 split, 4 to 6 weeks) so you get a clean A/B.

Step 6: Review on a schedule

• Day 3: search terms report filtered by match type = AI Max. Add negatives for anything irrelevant.
• Week 1: Assets report. Remove weak generated headlines. Check which final URLs are receiving expanded traffic.
• Week 2 to 4: compare CPA/ROAS between AI Max traffic and keyword traffic. Decide per ad group whether to keep search term matching on.

How to read AI Max reporting

The reporting is the real advantage AI Max has over Performance Max, so use it:

• Search terms report: filter the Match type column for "AI Max." Google also labels whether the match came from broad match expansion or keywordless matching. If you use ValueTrack, `{matchtype}` returns `a` for AI Max traffic, which lets you segment it in analytics.
• Keywords report: summary rows show how much of the campaign's volume came from AI Max versus your keywords.
• Combined search journey view: since late 2026 Google shows headline, landing URL, and search term together, so you can see exactly what Google matched, what copy it wrote, and where it sent the click, in one row. This is the fastest way to catch a bad expansion.
• Assets report: generated headlines and descriptions are labeled as automatically created, with performance ratings.

Doing the same thing faster with AI

Once you've done the manual pass, most of the ongoing work is reading reports and writing lists. That's exactly what Claude or ChatGPT are good at. Here are workflows you can copy.

Workflow 1: Audit AI Max search terms in five minutes

Export the search terms report (filter: match type = AI Max, last 14 days) as CSV. Then paste it into Claude or ChatGPT with this prompt:

```

You are a Google Ads specialist auditing AI Max search terms for an online store that sells [what you sell]. Our target CPA is $[X] / target ROAS is [Y].

Attached is a CSV with columns: search term, match type, impressions, clicks, cost, conversions, conversion value.

1. Group the search terms into themes (e.g., competitor brands, informational, wrong product, irrelevant, good fit).

2. For each theme, total the cost and conversions.

3. Return a list of negative keywords to add, in phrase match, grouped by theme, with one line explaining why.

4. Return a separate list of high-performing terms that we don't have as keywords and should add as exact match keywords in their own ad group.

5. Flag any terms where cost > 3x target CPA with zero conversions.

Keep the output in plain lists I can paste into Google Ads Editor.

```

You'll get a negatives list and a "graduate these to exact match" list in one shot. Review both; the model doesn't know your margins or your legal constraints unless you tell it.

Workflow 2: Write your brand exclusions and text guidelines once

```

I'm setting up Google Ads AI Max for [brand], which sells [category] in [country].

List: (a) 15 to 25 competitor and marketplace brands I should add as brand exclusions,

(b) up to 25 terms to exclude from AI-generated ad text (legal, pricing claims, and words that don't fit our premium positioning),

(c) up to 40 messaging restrictions phrased as short rules (e.g., "Never mention discounts or percentages off").

Our brand voice: [two sentences]. Things we never claim: [list].

```

Workflow 3: Build a URL exclusion list from your sitemap

Paste your sitemap URL list (Shopify exposes `/sitemap.xml`) and ask:

```

From this list of URLs, return only the ones I should EXCLUDE from Google Ads final URL expansion: blog posts, policy pages, account pages, collections with fewer than 3 products, out-of-stock pages, and anything without a purchase action. Output as one URL per line.

```

Workflow 4: Automate the weekly check with Claude Code

If you're comfortable with a terminal, ask Claude Code to write a Google Ads Script that runs weekly, pulls search terms with match type AI Max, flags anything over a cost threshold with zero conversions, and writes the list to a Google Sheet (and optionally emails you). A prompt like this is enough to get a working first draft:

```

Write a Google Ads Script (JavaScript, runs in the Google Ads Scripts UI) that: queries the search_term_view via GAQL for the last 7 days where the match type is AI Max; keeps rows with cost > 25 and conversions = 0; appends them to a Google Sheet (I'll provide the URL) with date, campaign, ad group, search term, cost, clicks; and emails me a summary. Include comments explaining each step.

```

Test it in preview mode first. Scripts can add negatives automatically too, but keep a human in the loop for the first month.

The ecommerce catch: AI Max optimizes for conversions, not profit

Everything above assumes the conversion signal you give Google is the right one. For most Shopify stores it isn't. The default purchase conversion sends revenue. AI Max then hunts for queries that produce revenue, which pushes it toward your cheapest, lowest-margin products, heavy discount searches, and items with expensive shipping, because those convert easily.

The fix is to send profit per order (price excluding tax minus product cost, shipping, payment fees) as the conversion value and bid on Maximize conversion value with a target ROAS that is actually a target POAS (profit on ad spend). Once you do that, search term matching and final URL expansion start finding queries that make you money instead of queries that make you busy. You can build this with a GA4 custom event plus a Google Ads offline conversion import, or use a tool that does it for you.

Alternatives to Google Ads AI Max

AI Max isn't the only way to get reach and automation in Search. Weigh these fairly:

Keyword-level broad match with Smart Bidding

• Pros: still available, still uses your keyword as a seed so expansion stays closer to intent, full search term visibility.
• Cons: no keywordless matching, so you'll miss long-tail queries that don't resemble any seed; no text customization.
• Best for: accounts that want to expand gradually and keep every query tied to a keyword.

Performance Max

• Pros: covers Shopping, YouTube, Display and more from one campaign; strong for ecommerce with a good Merchant Center feed.
• Cons: far less reporting, you can't see search terms at the keyword level, and it will cannibalize brand traffic unless you exclude it.
• Best for: stores that want volume across channels and already have Search covered.

Microsoft Advertising AI Max

• Pros: same three components, same mental model, usually cheaper CPCs, Bing traffic skews older and higher-income.
• Cons: smaller volume; controls are newer and less mature than Google's.
• Best for: anyone already importing campaigns from Google Ads.

Third-party search term automation (Optmyzr, Adalysis)

• Pros: n-gram analysis, automated negative rules, and alerts that work across Google and Microsoft; good at catching AI Max waste at scale.
• Cons: monthly subscription; still relies on the conversion data you feed Google.
• Best for: agencies and accounts with many campaigns.

Doing nothing (staying on exact/phrase)

• Pros: maximum control, predictable CPA.
• Cons: Google is migrating legacy structures into AI Max anyway, and exact match has been loosened over the years. You'll keep paying higher CPCs for shrinking volume.

How savvy.so does it

savvy.so plugs into your Shopify store and calculates the real profit of every order: price without tax minus product cost, shipping, payment fees and other costs. It sends that profit as the conversion value to Google Ads, Meta and GA4, so AI Max's Smart Bidding looks for margin instead of revenue, and shows you POAS by campaign and channel so you can judge whether AI Max traffic is actually profitable. The same profit data breaks down by product and category, which helps you decide which pages to exclude from final URL expansion. There's a free trial for Shopify stores.

FAQ