Google adds AI Max reporting, two years after advertisers called it a black box

Google is adding a Google Ads report that shows which search term, ad asset and landing page actually earned a conversion inside AI Max campaigns, the company said on September 23, alongside an expansion of its AI Brief guardrails feature from English into seven more languages. Neither has a ship date yet — Google says only that the reporting view is coming "later this year."

What Google is shipping

The September 23 post describes two changes. AI Brief, the natural-language guardrail tool that lets advertisers tell AI Max what to say and what to avoid, moves out of English-only closed beta into Dutch, French, German, Italian, Japanese, Portuguese and Spanish. Separately, Google Ads will get what the company calls "a single, unified view of the Search ads journey" — one report showing the search term that triggered an AI Max ad, the creative asset the user saw, and where they landed on the advertiser's site. Google has not published a screenshot, a rollout date, or which account types get it first.

Both changes land squarely on software marketers running paid Search. AI Brief's language expansion matters most to any SaaS vendor buying keywords outside English-speaking markets, who until now had no way to set messaging and matching guardrails on AI Max at all — a gap that mattered because the automation has a documented history of drifting off-message without them. The reporting change matters to everyone running AI Max, English-language accounts included, because it targets the specific complaint practitioners have made for two years: that the feature's own numbers can't be trusted to say what actually drove a result.

Why advertisers have been asking for this since 2024

AI Max for Search launched in May 2025 already marketed as giving advertisers "the granular controls and reporting transparency you need." Google's own launch post credited it with a typical 14% lift in conversions or conversion value, rising to 27% for accounts still running mostly exact and phrase match.

The transparency didn't hold up under inspection. In December 2024, Adalysis co-founder Brad Geddes documented that AI Max was claiming credit for search terms that identical exact-match keywords were already capturing, that some "keywordless" matches had no keyword attribution at all, and that a search for "preschool near me" had once matched to a pricing ad group instead of the advertiser's own location-targeted one. Google's Ginny Marvin, then the company's Ads Product Liaison, replied on LinkedIn that the mismatch came from matching on inferred intent rather than the literal text typed into the search box.

"We're increasingly determining relevance by inferred intent (like with Lens or AI Overviews) versus just the raw text query."

Marvin said Google would update its documentation the following quarter. A year later, the underlying problem was still unresolved. A December 2025 Adalysis post found AI Max still books traffic against both a keyword's original match type and broad match at once, so, in Geddes' words, "you can no longer evaluate how each match type is performing." The same post found AI Max routinely matching brand terms to competitor searches and non-brand terms to brand ones inside the same account.

By April 2026, the measurement problem had outgrown any one platform quirk. Writing in Search Engine Journal, Brooke Osmundson, director of growth marketing at Smith Micro Software, argued that PPC's old cause-and-effect measurement model "is breaking down in real time" as AI Max and similar automated campaign types generate conversions from queries, assets and placements advertisers never explicitly chose. Her recommendation was to stop reporting by literal keyword text and start grouping performance by intent cluster instead. Then, on September 1, 2026, Google began auto-upgrading Search campaigns that had only opted into automatically created assets or campaign-level broad match, switching all of AI Max on for them whether the advertiser asked for it or not — pulling more spend into the same reporting blind spot the September 23 fix is meant to close.

What to check in your own account

• Manually create broad-match versions of your strongest exact and phrase keywords, giving AI Max's traffic a keyword to attach to that you can still isolate and compare — one workaround analysts use to keep match-type data readable ahead of Google's fix.
• Build negative keyword lists that separate brand, non-brand and competitor terms; Adalysis found AI Max blurring all three inside the same account.
• Check whether your account was swept into the September 1 auto-upgrade, and compare match-type spend before and after.
• Report AI Max performance by intent cluster, not keyword text, and hold out a non-AI-Max campaign as a comparison baseline.
• When the new reporting view ships, reconcile it against your own conversion logs before treating it as the account's source of truth — Google has revised its explanation of AI Max matching at least once already.