What changed
Google AI Max for Search is now generally available. It was introduced in beta in late 2025, reached GA in early 2026, and got a broader rollout announcement at Google Marketing Live on May 20, 2026, along with previews of AI Max for Shopping and AI Max for Travel.
The feature is a single campaign-level toggle. Flipping it on activates three capabilities simultaneously: search term matching expansion (Google's AI matches your ads to queries beyond your keyword list, using query understanding rather than strict match types), text customization (Gemini generates additional headlines and descriptions tailored to each query and, where relevant, the landing page), and final URL expansion (the system dynamically routes users to the page it judges most relevant, not necessarily the URL you specified in the ad).
Google's reported benchmark: campaigns using the full AI Max suite see an average 7% more conversions or conversion value at a similar CPA/ROAS compared to using search term matching alone. The GA release also shipped more granular controls, brand include/exclude, location-of-interest targeting, URL controls to constrain final URL expansion, and text guidelines to bound what Gemini can write.
Who it affects
If you run Search campaigns with curated keyword lists, AI Max changes the fundamental contract between your keyword structure and who actually sees your ads. That affects almost everyone, but the stakes are highest for accounts where brand terms are competitively sensitive, for multi-page sites where landing page relevance drives conversion rate, for regulated verticals where copy must be reviewed before it runs, and for agencies managing accounts where the client has approval authority over ad text.
Why it matters
The mechanism here is a shift in where control lives. Traditional Search puts the advertiser in control of the query trigger, you pick keywords, set match types, and the auction matches from there. AI Max moves that trigger upstream into Google's query understanding model, with the advertiser managing guardrails instead of inputs.
That's not inherently bad. Broad match with Smart Bidding (Google's automated bid strategy tied to conversion signals) already demonstrated that loosening keyword constraints can surface converting queries humans wouldn't have thought to add. AI Max formalizes and extends that logic, adding copy and URL adaptation on top.
The 7% conversion lift is plausible given that framing, you're reaching more relevant queries and sending users to better-matched pages. But that number is an average across campaigns that presumably had controls configured correctly. For accounts that go live without tightening brand exclusions or URL rules, the conversion volume might rise while CPA also rises, because spend is distributing across a wider query set that includes noise.
The text customization piece carries its own risk. Gemini generating headlines means copy is running that no human approved. Text guidelines let you constrain this, but they require active configuration. An account that enables AI Max and treats the text guidelines as optional is effectively running an open brief on its own ad copy.
The play
-
Clone before you commit. Run AI Max in a copied campaign against the original for at least four weeks with equal budget. Watch CPA or ROAS; set a pass bar at your 30-day account average plus a 10% tolerance. Don't consolidate until the test is clean.
-
Configure controls before launch, not after. Set brand exclusions for any terms you don't want to own in the auction. Add URL controls to restrict final URL expansion to your approved landing page set. Draft text guidelines that explicitly exclude claims, formats, or tones your brand doesn't use. Do this before the campaign goes live, retrofitting controls on a live campaign means a window where the AI runs unconstrained.
-
Pull the search terms report at 7 days. AI Max will surface queries you've never seen. Flag anything irrelevant and add negatives immediately. Treat this as a standing weekly task for the first 60 days, not a one-time audit.
-
Review AI-generated copy at 30 days. Use the asset report to see which Gemini-written headlines and descriptions are running. Confirm they're within your guidelines. If off-brand copy appears, tighten the text guidelines and check whether those assets correlate with worse conversion rates, if they do, that's a signal the model needs tighter constraints, not just a compliance issue.
-
Prepare for AI Max in Shopping. Google previewed AI Max for Shopping at GML 2026. If Shopping drives meaningful revenue for your account, start thinking now about what query expansion and dynamic content generation would mean for your feed quality and product page relevance, the same control questions will apply.
Watch-outs
Final URL expansion is the highest-risk element for most accounts. If your site has category pages, blog content, or policy pages that could technically satisfy a query but have no conversion intent, the AI can route budget there. URL controls exist to prevent this, but they require you to explicitly define the approved URL set, the default is expansive.
Text customization can produce factually accurate but legally problematic copy in regulated categories (finance, health, legal). Don't assume Google's content policies catch everything your compliance team would flag.
On the query expansion side, watch for competitor brand queries appearing in your search terms report. AI Max may match to them if it judges your product relevant. That has bid and budget implications, and in some categories, legal ones.
The WhyItWon angle
AI Max means the ad your audience actually sees, the query that triggered it, the headline Gemini wrote, the page it landed on, is increasingly generated at runtime rather than specified in advance. Your keyword list and your approved copy are now starting points, not final outputs. That raises the stakes on knowing, before you spend, which creative signals actually drive conversion for your account and your audience. WhyItWon reads your existing ads, your competitors', and your customers' language to score what's likely to win before budget commits, which matters more, not less, when the platform is adapting your creative on the fly.
Google Ads
