What is the AI Max Search terms and ad combinations report?

It is a report for AI Max for Search campaigns that, for the first time, puts the search query, the ad combination shown (up to three headline assets concatenated and separated by pipes), the landing page reached, and the campaign and ad group all in one row. It sits under the Search terms dropdown in the Google Ads interface, where you pick “Search terms and ad combinations” instead of the classic search terms view. Crucially it carries a match-source label that tells you why AI Max served each query — whether it came from AI Max expanded matching or from a landing-page/URL inclusion. It began appearing in select accounts around mid-October 2026 and, at the time of writing, is not yet generally available, so if you do not see the option yet it has not reached your account.

Why look at the landing page instead of just adding negatives?

Because in an automated campaign the landing page is often the thing pulling the junk queries, not an accident downstream of them. AI Max reads your final URLs and page content to decide which searches to expand into, so one weak or off-topic page can act as a magnet for a whole family of irrelevant queries. If you only negate the terms it drew, you are mopping the floor with the tap running: the page keeps attracting new variants faster than you can list them. Finding the one or two pages responsible and fixing or excluding them removes the cause rather than the symptoms, and it is far less work than adding negatives one term at a time forever.

Can I sum the numbers in this report against the classic search terms report?

No. The combination view overlaps the classic search terms report, so adding metrics across the two double-counts clicks, cost and conversions. Google's own guidance is to never sum across these views and to reconcile against campaign totals instead. Treat the new report as a diagnostic lens for seeing which page and ad combination a query landed on — not as an additional pile of spend on top of what your campaign already reports. The moment you start adding the two together your waste figures become fiction, and every prioritisation decision you make from them inherits that error.

How do I isolate just the AI Max expanded traffic in this report?

Filter on the match source. In the interface the report labels each row as an AI Max expanded match or a landing-page/URL inclusion; in the Google Ads API the equivalent is filtering the search term view on search_term_match_source, and the dedicated ai_max_search_term_ad_combination_view exposes the query, landing page and headline fields together. Filtering this way strips out the traffic your existing keywords were always going to match and leaves only the queries AI Max added on its own — which is the traffic actually worth auditing for waste, because it is the part you did not explicitly ask for.

Should I fix the page or add a negative keyword?

Match the remedy to the match source. If a bad query arrived through AI Max expanded matching, a negative keyword is the right tool — it tells the automation to stop expanding into that query space. If it arrived through a landing-page or URL inclusion, a negative does not address the cause, because the page itself is why AI Max thought the query was relevant; there the fix is to rewrite the page so its content stops signalling the wrong intent, or to exclude that URL from the inclusions AI Max is allowed to use. Reading the match-source column before you act is what stops you from adding a negative that will never fire against a page-driven problem.

AI Max: Find the Landing Page Bleeding Spend, Not the Term

For most of Performance Max and AI Max's existence the search terms report answered one question — what did we match on? — and left you to guess the rest. You could see that a junk query had spent money, but not which ad it triggered or which page it dumped the visitor on. The new Search terms and ad combinations report for AI Max, which began appearing in select accounts around 13 October 2026, closes that gap: the query, the headline combination shown, the landing page reached, and the campaign and ad group now sit in a single exportable row, each tagged with the match source that explains why AI Max served it.

That extra column — the landing page — changes what a search-terms audit can do. Until now the unit of waste was the term: you found a bad query and you negated it. But in an automated campaign the page is frequently the thing pulling the bad queries, and when it is, negating terms one at a time is mopping the floor with the tap running. This post is about using the new report the way the extra data makes possible: pivot the export by landing page, rank pages by how much wasted query volume each one concentrates, and fix the worst offenders at the page or URL level instead of chasing their symptoms through an ever-growing negative list.

What the unified report actually adds

The report lives under the Search terms dropdown in the Google Ads UI, where you now choose “Search terms and ad combinations” alongside the classic view. Each row carries the search query, the ad combination that served against it — up to three headline assets concatenated and separated by pipe characters — the final landing page, and the campaign and ad group. The field that matters most for diagnosing waste is the match-source label, which, per PPC Land's coverage, distinguishes AI Max expanded matches from landing-page and URL inclusions. That single column is the difference between knowing a query was wrong and knowing why AI Max thought it was right.

Before this, the landing page a query reached was invisible in the search terms report — you would have needed to cross-reference the asset and URL reports and reconstruct the join by hand, if it was recoverable at all. Having the page on the same row as the query is what makes a page-level audit a two-minute pivot rather than a data-engineering project. It is also the same transparency arc Performance Max went through when channel-level reporting arrived; if you ran that audit, the PMax spend-distribution audit is the direct ancestor of what follows, one level deeper.

Why the landing page is the unit of waste

AI Max decides which queries to expand into partly by reading your final URLs and the content on them. That means a single page can act as a magnet: if a product page talks loosely about a category, lists adjacent use-cases it does not actually serve, or carries stale copy about a discontinued line, AI Max will read that as permission to match a whole family of queries you never wanted. The waste is not a scatter of unrelated bad terms — it is a cluster, all pointing at the same URL, because the same page content generated all of them.

This is exactly why term-by-term negation loses the race. You negate “cheap”, the page draws “discount” and “budget” and “free”; you negate those, it draws “second hand” and “used”. The list grows, the page keeps producing variants, and your waste stays flat. Treating the page as the unit flips the economics: fix the one URL and the entire cluster of queries it was generating stops at the source. It is the same logic behind grouping patterns instead of listing single negatives, except the grouping principle here is not a word stem — it is the page the queries share.

Pivoting the export by landing page

Export the report and pivot it so the landing page is the row dimension, not the query. For each page, sum the cost and clicks and count the distinct queries it served, then sort pages by cost descending. What you are looking for is concentration: a small number of pages accounting for a large share of spend on queries you would not have chosen. A page that drew forty irrelevant queries for a meaningful slice of the campaign's cost is a far more valuable find than forty separate negatives, because it is one fix instead of forty, and it addresses the thing that will otherwise keep generating number forty-one.

Filter to the AI Max match source before you do this, so you are auditing only the traffic AI Max added rather than the queries your existing keywords were always going to match. In the interface that is the match-source column; via the API it is a filter on search_term_match_source, and the dedicated ai_max_search_term_ad_combination_view returns the query, landing page and headline fields together for exactly this purpose. The mechanics of separating AI Max traffic from your keyword traffic are the same ones covered in segmenting the AI Max search terms report; here you are simply carrying the landing page through the segmentation as well.

Do not sum this report against the classic one

One discipline before you trust any number you pull from here: the combination view overlaps the classic search terms report, so you must never add metrics across the two. Doing so double-counts clicks, cost and conversions, and every waste figure you compute downstream inherits the error. The report is a diagnostic lens for seeing which page and ad combination a query landed on — it is not an extra pile of spend sitting on top of what the campaign already reports. Reconcile to campaign totals, not to the sum of two overlapping views.

There is a second trap in the landing-page column specifically: the final URL recorded is not always the URL the user actually reached, because redirects and tracking templates can send them somewhere else. Google exposes a separate expanded-landing-page view for the URL that was genuinely served, so if a join keyed on the final URL looks wrong, that is the view to reconcile against. This is the same family of cross-view inconsistencies that the AI Max search-term attribution discrepancies post covers for conversion counts — the fix is always to understand which view answers which question, not to force two of them to agree.

Page fix or negative keyword? Read the match source

The match-source label is not decoration — it tells you which tool to reach for. A bad query that arrived through AI Max expanded matching is a negative-keyword problem: the automation chose to expand into that query space, and a negative is how you tell it to stop. A bad query that arrived through a landing-page or URL inclusion is a page problem: the page content is why AI Max thought the query was relevant in the first place, and a negative keyword does nothing to change that signal. Add a negative against a page-driven match and you have filed a term that may never fire while the page keeps drawing new variants around it.

So the decision rule is: for expanded-match waste, negate — grouping recurring patterns rather than listing single terms, and watching that you do not throttle good traffic, the risk covered in over-negating and throttling good traffic. For inclusion-driven waste, fix the page: rewrite the copy so it stops signalling the wrong intent, or exclude that URL from the set AI Max is allowed to pull from. This is the page-level analogue of distinguishing the match sources in the first place, the subject of reading the AI Max source column — except here you are routing each source to a different remedy rather than just interpreting it.

The weekly workflow in your account

Put together, the routine is short. Open the Search terms and ad combinations report, filter to the AI Max match source, and export. Pivot by landing page and sort by cost. Take the top two or three pages by wasted spend and, for each, read the match-source split of the queries it drew: the expanded-match queries go into a grouped negative, and if the bulk arrived via inclusions, the page itself goes on the list to rewrite or exclude. Reconcile your totals to the campaign, never to the sum of the two reports, and note the pages you have actioned so next week you are looking for new offenders rather than re-reading the same ones.

Fold this into the standing cadence rather than treating it as a one-off. The broader 2026 search hygiene audit playbook is where this pivot belongs — a weekly or fortnightly pass once the campaign is past its ramp. The report is new and still rolling out, so the field names and exact labels may shift before it reaches general availability; the method does not. Whatever Google calls the column, the move is the same: let the landing-page dimension tell you where the waste concentrates, and fix the cause instead of chasing the symptoms.

For how to read the match-source column itself — broad expansion versus keywordless and inclusion sources — see the AI Max source column explained. For when the right response is negating the term because the ad shown mismatched the query, see negating on creative mismatch. And for the wider spend-distribution view one level up, the PMax channel-level reporting audit is the companion piece.