If you buy media for a living, the real story behind Alphabet's massive $119.8 billion Q2 earnings was clear — AI isn't just a "nice-to-have" campaign feature anymore. Google is rewriting the entire advertising stack around it, and the days of tightly controlled keyword match types are ending.
Here is what Google's AI odyssey means for your dashboards, and the playbook to protect your budgets from algorithm bloat.
AI Max is a Trojan Horse for Broad Match
There has been a lot of confusion regarding AI Max, with many buyers assuming it is a replacement for Performance Max.
Google's updated help documentation explicitly clarifies that it isn't a new campaign type at all. It is an optimization layer baked directly into your existing Search campaigns.
By opting in, you are giving Google permission to use keywordless technology and dynamic URL expansion to match your ads to messy, conversational queries. Google claims advertisers see a 15% lift in conversions at a similar ROAS.
While unlocking new long-tail demand sounds great, experienced operators know what happens when you hand Google the keys to broad-match expansion: they monetize their garbage inventory on your dime.
The Siren Song of Generative Search Ad Placements
We all knew it was coming. Now that Gemini is handling billions of complex, conversational queries, Google is actively testing ads inside AI Mode.
The platform is rolling out "Highlighted Answers" — clearly labeled sponsored links injected directly into AI-generated list responses.
For search marketers, Generative Engine Optimization is no longer an SEO buzzword; it is a paid media necessity.
If you aren't feeding Google the precise structural data it needs to synthesize these answers, your competitors are going to steal your high-intent traffic before the user ever sees a traditional organic link.
The Funnel is Collapsing
Google is actively shrinking the distance between top-of-funnel discovery and bottom-of-funnel checkout.
With the rollout of Universal Cart and the Universal Commerce Protocol, users can now add products to a cart from YouTube, Gemini, or Gmail, and checkout in one unified flow. Marketing that stops at the click is becoming obsolete.
4 Takeaways to Protect Your Margins in Q3
Google isn't selling ad space anymore; they are selling intent matching. To survive the shift, you have to stop complaining about the algorithm and start fixing your inputs.
Before you turn AI Max loose — or before Google auto-upgrades you in September — execute this checklist:
1. Set Your Pre-Flight Guardrails
If you turn on AI Max in a non-brand generic campaign, it will immediately start bidding on your own brand terms to claim that algorithmic lift. Set hard campaign-level brand exclusions on day one.
Next, audit your Final URL Expansion (FUE). If you have a high-converting dedicated landing page, heavily restrict FUE with URL exclusions so the AI doesn't start dumping paid traffic onto your standard "About Us" page.
2. The Server-Side Tracking Mandate
Browser tracking is currently missing 30-60% of your actual conversions due to ad blockers and iOS updates.
If you hand the keys to AI Max with a broken pixel, the algorithm will optimize toward a biased, incomplete sample — and scale that mistake across your entire budget.
Before you launch, you must have Server-to-Server tracking (like Enhanced Conversions) fully operational. The buyer with the most accurate first-party data wins the auction.
3. Check the AI’s Recipe
Google wants you to look at high-level ROAS, but operators need to look at the query level.
Pull the ai_max_search_term_ad_combination_view report weekly. This shows the exact "recipe" the AI used — the user's search term, the dynamically selected headline, and the specific landing page it chose.
Use this to identify where the AI is mismatched so you can add negative keywords and URL exclusions.
4. Restructure Your Feed for Universal Commerce
Google's AI now treats your Google Merchant Center feed as a deep database, extracting attributes to dynamically rewrite your product titles to match conversational intent — which means short, generic feed titles won't move the needle.
Enrich your feed with hyper-specific attributes (e.g., "waterproof," "stability-focused," "compatible with X").
The more raw data points you give the AI, the better it can dynamically match your products to the messy, long-tail queries happening in AI Mode.
The Algorithm Isn't Magic, It's Just Hungry
The media buyers who win this quarter won't be the ones meticulously tweaking exact-match bids. They will be the ones who feed the algorithm the cleanest data, set ruthless brand guardrails, and force the AI to work for them.
Join the AdLeaks community now for unique insights on Google Ads, AI, and more!
Related Posts:
