BLOG / BLOG POST

Google AI Max for B2B: Turn It On, Then Fence It In

AI Max usually wins on volume and cost. What it quietly takes from you is control over which page the click lands on and which searches you show up for.

Google's AI Max for Search is not another campaign type you can ignore. It is a set of features that live inside the search campaigns you already run, and in most accounts it produces more conversions at a lower cost within a few weeks of being enabled.

That is the honest starting position. The problem in B2B is that the same automation that finds new profitable queries will also route a $60 click for an enterprise buying question to your careers page, in a country you do not sell to, under a headline it wrote itself.

So the decision is not whether to use it. It is which fences you build before you turn it on.

What AI Max actually changes in a search campaign

AI Max replaces keyword-based matching with intent-based matching, then takes over creative and landing page selection unless you tell it not to.

Three features do most of the work, and Google documents them together:

Search term matching finds queries with no keyword in your account, using your landing pages, existing keywords, and assets as context.

Text customization writes new headlines and descriptions tailored to the query, drawing from your site and your existing assets.

Final URL expansion picks whichever page on your domain it predicts will convert best for that query, not the URL you set in the ad.

Each of those is a control you used to hold. Keywords decided which searches you paid for. Your RSA assets decided what the ad said. The final URL decided where the click went. AI Max makes all three of those a prediction instead of a setting.

For a B2B account with a $200 cost per lead and a six month sales cycle, that matters more than it does for an ecommerce store selling running shoes. A bad ecommerce click costs a dollar. A bad enterprise click costs a lunch.

Ready for paid ads that pay off?

Book your free audit

Why B2B accounts should still test it

The reason to run it is that keyword coverage in serious B2B categories is worse than most teams admit.

Real buyers describe problems, not products. They search in full sentences, they use internal jargon from their own industry, and a meaningful share of their searches have never been typed before. No keyword list built from a planner tool covers that tail, and exact-match discipline leaves it on the table.

That tail is where AI Max earns its keep. It is also where the waste lives, which is the entire reason the controls exist.

The pattern we keep seeing in accounts that enable it is the same shape: conversion volume climbs, cost per conversion drops, and the search terms report fills with queries nobody would have written down. Some of those queries are excellent. A handful are competitor names, jobseeker questions, and free-tool hunters. The accounts that come out ahead are the ones that read that report every week for the first month.

The four fences to set before you enable it

Lock down where the traffic can land

Final URL expansion is on by default when you opt into AI Max, and it is the setting with the most downside in B2B.

Your site probably has pages that should never receive paid clicks. Careers. Support docs. Old blog posts. A pricing page you are in the middle of rewriting. A partner directory. Left alone, expansion treats all of it as fair game if it predicts a match.

Use URL exclusions to block the sections that should stay out of paid acquisition, and URL inclusions to point it at high-intent pages it has not discovered on its own.

One detail that catches teams by surprise: Google states that pinned RSA assets are not respected when final URL expansion selects a different page. If you rely on pinning to keep a compliance line or a specific product name in position one, expansion will break that. Either turn expansion off in that campaign or drop the pin and solve it another way.

Tell it which brands it can appear next to

Brand controls work at both campaign and ad group level, and they run in two directions. You can require that your ads only show for searches associated with specific brands, or you can block your ads from appearing alongside brands you do not want to be near.

In practice B2B accounts use exclusions more than inclusions. Competitors you have a legal reason to avoid naming, partners whose terms restrict you, and resellers whose queries you do not want to buy all belong in that exclusion list on day one, not after somebody screenshots your ad.

Set locations of interest at the ad group level

Geography is where AI Max leaks fastest, because intent-based matching does not care about your service map.

AI Max adds locations of interest, an ad group level control that targets people based on the geography their search implies rather than only where they physically are. That is genuinely useful for a company selling to one region from another, and it is a different lever from the campaign level presence versus interest setting that already trips up most accounts. If you have not sorted that one out yet, start there, because layering AI Max on top of a loose geo setting compounds the problem.

The specific failure to watch for is traffic from countries you do not serve arriving through keywordless matches. Exclude those markets explicitly. Do not assume a country you never added is a country you never bought.

Keep your negative keyword workflow running

Negative keywords still function inside AI Max. This is the part people get wrong when they hear the word keywordless: your negatives remain in force, and they are now doing more work than before, because there is no keyword list acting as a first filter.

That means the weekly negatives habit is not optional anymore, it is the primary steering mechanism. The search terms report is your keyword list now. Read it, cut what does not belong, and keep the exclusions campaign-specific so you are not blocking a term that qualifies in one context and wastes money in another.

Point it at a qualified conversion or do not bother

AI Max optimizes toward whatever you tell it is a conversion, and it will find volume with unnerving efficiency.

Feed it raw form fills and it will deliver more form fills, including from students, consultants, and people who want your free calculator. That is not the automation failing. That is the automation doing exactly what you asked.

Before you enable it, the account needs a conversion signal tied to something a salesperson would recognize as a real opportunity. A qualified stage from the CRM. A held meeting. A sales-accepted lead. If you only have a few of those per month, use volume-based bidding on the best qualified event you have rather than jumping to value-based bidding on a signal you do not trust yet.

Enabling broad automated matching on top of a form-fill goal is how accounts end up with a great-looking dashboard and a sales team that stops answering marketing's Slack messages.

What to check in the first three weeks

Set a recurring 20 minute slot and look at four things in order:

1.) Search terms. Which new queries did it find, and would you have bid on them deliberately? Add the good ones as keywords where it helps, negate the rest.

2.) Landing pages chosen. AI Max reports which URLs expansion selected. If a support page or a careers page is in that list, your exclusions are incomplete.

3.) Asset text. Read the auto-generated headlines. B2B categories have specific language, and generated copy sometimes gets a product claim subtly wrong in a way legal will care about.

4.) Lead quality by source, not lead count. Compare qualified rate before and after, not total conversions. Volume almost always goes up. The question is whether the qualified rate held.

If the qualified rate drops and stays down for three weeks with the fences in place, turn expansion off first, not AI Max entirely. Expansion is the feature most likely to be responsible in a B2B account with a large, mixed website.

The uncomfortable part

AI Max is Google removing the last few settings that let a mediocre account structure survive on operator skill. What is left to be good at is the input quality: your conversion definition, your landing pages, your exclusions, and how carefully you read what the system did last week.

Teams that treat those four things as their job get more from AI Max than they got from any keyword build. Teams that flip it on because a Google rep said it lifts conversions get a nice chart and a worse pipeline.

Start with one campaign, not the whole account. Turn on search term matching and text customization, leave final URL expansion off for the first two weeks, and only enable it once your URL exclusions are written down. If you want a second set of eyes on the controls before you touch the account, that is exactly what our B2B Google Ads work is for.

share this article

Peter Guba

Author

Peter Guba

CEO of Profit Mill

About Peter

Keep up with the latest insights

Want to see what a performance-driven Google Ads strategy can do for your business?