Google killed search term transparency the same week it made AI Max mandatory

Google killed search term transparency the same week it made AI Max mandatory. That timing tells you everything.

Starting September 2026, Dynamic Search Ads are retired. AI Max is the replacement. You do not get to vote.

At the same moment, Google changed search term reporting so only queries generating 100+ impressions are visible. The long tail, where most B2B and niche campaigns live, goes dark.

It gets worse. The "search terms" you do see in AI Max are not actual queries users typed. They are AI-built intent approximations. You are auditing a synthetic representation of your own traffic.

Why each of these matters

The impression threshold for search-term visibility sounds reasonable until you run the numbers on a real account. Most B2B Google Ads accounts have long tails where 60 to 80 percent of conversions come from queries that individually generate fewer than 100 impressions in a 30-day window. Those queries are now invisible in reporting. You can see they exist in aggregate, in the conversion count, but you cannot see the words.

The negative-keyword strategy that depended on weekly search-terms review just stopped being possible at the same depth. You can still negate the queries that cross the visibility threshold. You cannot negate the long tail, because you cannot see it.

The synthetic search-term reporting in AI Max is the more material change. When Google says the report shows "user intent that matched your ad," that means the AI inferred the intent and grouped queries by its own clustering, not by the actual strings users typed. You are looking at the model's interpretation of what users asked, not the data itself.

The independent verification step that used to ground every paid-search audit, "show me the actual queries that drove this conversion," no longer returns the truth. It returns the platform's interpretation of the truth.

What you can still see, what you cannot

You can still see:

• Aggregate conversion data at the campaign and ad-group level
• Impression share, CPC trends, conversion rate
• High-volume search terms above the threshold
• Audience-level reporting where you have first-party data

You cannot reliably see:

• Long-tail queries below the impression threshold
• The literal strings AI Max matched to your ads
• Whether AI Max is cannibalising your branded search at inflated CPCs (without doing the work below)
• The actual incremental lift of AI Max vs the campaigns it replaces

The only edge left

If the algorithm has to learn from your data, and your first-party data is cleaner than your competitor's, the black box works harder for you. It is a thin moat, but it is the one that remains.

What clean conversion signal actually looks like:

• Enhanced Conversions for Leads correctly implemented across every lead form
• Offline conversion import from your CRM, with realistic conversion values that reflect actual closed revenue
• Conversion Value Rules tuned for the audience segments that produce the highest lifetime value
• Negative-keyword lists at the account level (not just campaign level) that exclude irrelevant intent before the auction happens
• Conversion event hygiene, no spam form submissions counting as conversions, no double-counting from view-through pixels

Most accounts have one or two of these. Almost none have all of them.

What to actually do

• Audit your conversion definitions this week. If "conversion" still means "form submission" rather than "qualified lead", AI Max is being trained on noise that will become harder to undo once the platform locks in the bidding strategy.
• Add account-level negative keyword lists for your brand terms specifically. AI Max can intercept branded queries at inflated CPCs if you let it.
• Implement offline conversion import if you have not. The 60-90 day project is worth more in the AI Max era than it was in the manual era.
• Export your full historical search-term reports before the threshold change locks them out. They will inform your negative-keyword strategy long after the live reporting stops showing them.
• Build an independent audit layer. Server-side tagging, a CRM-aligned conversion definition, periodic comparison of platform-reported revenue vs actual revenue. The platform's reporting is now an opinion. Your independent measurement is the result.

So here is the question: what signal are you actually feeding AI Max? Not what you plan to feed it. What is live right now.

If you want a view on whether your conversion infrastructure is ready for AI Max becoming the default, book a free audit and we will tell you where the signal is breaking before the algorithm builds bidding habits around the wrong data.