Last updated: February 10, 2026
AI Max for Search campaigns has gone from an experimental beta to Google's fastest-growing Search ads product in under a year. Announced at Google Marketing Live in May 2025, rolled out globally in beta through Q3 2025, and now available across Google Ads, Google Ads Editor, Search Ads 360, and the Google Ads API, AI Max represents the most significant change to how Search campaigns work since responsive search ads replaced expanded text ads.
If you tested AI Max in late 2025 and moved on, it is time to revisit. The feature has evolved substantially. One-click experiments, text guidelines for brand-safe creative, expanded reporting transparency, and integration with Smart Bidding Exploration have all been added or improved since the initial launch. The performance data has matured too. We now have enough real-world results across thousands of accounts to separate what actually works from what Google's marketing claims.
This guide covers everything you need to know about AI Max in February 2026: what it actually does, what changed since launch, the real performance data (not just Google's cherry-picked case studies), the prerequisites your account must meet before testing, best practices for implementation, common failure points, and how AI Max interacts with autonomous campaign management.
The name "AI Max" has caused significant confusion since launch. It sounds like a new campaign type, similar to Performance Max. It is not. AI Max is a suite of AI-powered features that you enable as an optional enhancement layer on your existing Search campaigns. You toggle it on, and three capabilities activate simultaneously.
This is the core feature. AI Max expands your campaign's reach beyond your existing keyword list by using broad match and keywordless technology to find relevant queries based on your keywords, landing pages, ad assets, and URLs. If you bid on "emergency plumber," AI Max might also serve your ads for "burst pipe help near me" or "water leak repair urgent" without you ever adding those keywords manually.
This is not fundamentally new technology. It combines the logic of broad match keywords and Dynamic Search Ads into a single, more sophisticated system. What makes it different is the integration: AI Max uses signals from your entire account (keywords, landing pages, conversion data, audience behavior) to predict which queries will convert, rather than relying purely on keyword semantics.
Your existing keyword priority rules still apply. Exact match keywords take priority over AI Max matches. Phrase match and broad match keywords compete with AI Max on equal footing, with Google serving whichever it predicts will perform better. This means AI Max primarily discovers queries that your current keyword list does not cover, rather than replacing your keywords entirely.
AI Max's second component dynamically generates headlines and descriptions using Google's Gemini model. The system analyzes your landing pages, existing ad copy, and the user's query to create ad text that matches the specific search intent. If a user searches "commercial roofing contractor warranty," AI Max might generate a headline like "Commercial Roofing with 10-Year Warranty" pulled from your warranty page, even if you never wrote that headline manually.
This feature was previously called "automatically created assets" and existed in a less sophisticated form before AI Max bundled it into the suite. The improvement is in the quality and relevance of generated text, which has gotten noticeably better through Google's Gemini integration.
Final URL expansion is also enabled by default under asset optimization. This allows AI Max to send users to whichever page on your site it deems most relevant, even if it is not the landing page you specified in the ad group. This is powerful when it works (matching users to exactly the right page) and dangerous when it does not (sending users to irrelevant pages or pages not optimized for conversion).
The newest addition to the AI Max suite. Locations of Interest allows AI Max to target users based on geographic intent at the ad group level. If someone in Dallas searches "plumber in Austin," AI Max can serve your Austin plumbing ad because it recognizes the geographic intent in the query, even though the user is physically in Dallas. This is particularly valuable for businesses that serve multiple markets and for travel, hospitality, and real estate advertisers.
AI Max in February 2026 is meaningfully different from AI Max at its September 2025 global launch. Here are the updates that matter.
Google built experimentation directly into the AI Max activation flow. When you enable AI Max on a campaign, you can choose to run an experiment rather than activating it outright. The system creates a 50/50 traffic split between your campaign with AI Max enabled and the same campaign with AI Max disabled. Google handles all the statistical analysis, tracking conversion lift, cost efficiency, and confidence intervals. You get clear, comparative results in the Experiments section.
This was arguably the most important post-launch update because it removed the biggest adoption barrier: risk. Previously, testing AI Max meant either enabling it and hoping for the best or duplicating campaigns and manually splitting traffic. Now, you can run a scientifically valid test with a single toggle. Experiments can be set to auto-apply when results reach statistical significance, or you can review results manually and decide whether to keep AI Max enabled.
Text guidelines address the most legitimate concern about AI-generated ad copy: brand safety and voice consistency. The feature works through two mechanisms. Term exclusions let you specify words and phrases the AI should never use. If "cheap," "bargain," or "discount" do not align with your premium positioning, you exclude them. Brand guidance lets you provide natural language instructions about your brand voice, messaging priorities, and content requirements that steer the AI's creative output.
Text guidelines were announced in September 2025, began rolling out in limited beta in fall 2025, and are expanding to all advertisers through Q1 2026. Google has confirmed that text guidelines will also apply to Performance Max campaigns, creating a unified brand control framework across both campaign types.
Smart Bidding Exploration is a separate feature that pairs naturally with AI Max. It allows Google to temporarily lower your effective ROAS target (within an acceptable range you define) to explore new traffic that would not have qualified under your standard target. When combined with AI Max's expanded query matching, Smart Bidding Exploration lets both the queries and the bidding expand simultaneously.
Google's internal data shows campaigns using AI Max with Smart Bidding Exploration see an 18% increase in unique search query categories with conversions and a 19% increase in overall conversions. These numbers come from Google, so they should be taken with appropriate context, but the directional signal is clear: the combination finds incremental volume.
AI Max now includes match-type indicators in search terms and keywords reports that specifically identify which queries were matched through AI Max versus traditional keyword matching. Combined views show the matched term, the headlines that served, and the landing URLs that were used. This transparency is a significant improvement over the initial launch, where it was difficult to separate AI Max performance from standard keyword performance.
AI Max is now fully supported in Google Ads Editor and through the Google Ads API. At launch, it was only available through the Google Ads web interface, which created friction for agencies and large advertisers managing campaigns at scale. Full API support means AI Max can be enabled, configured, and monitored programmatically.
Search ads served through AI Max-enabled campaigns can now appear in Google's AI Overviews (the AI-generated summaries at the top of search results) on both desktop and mobile across multiple global markets. Google is also testing ads in AI Mode, where users ask deeper, more complex questions. While advertisers cannot yet bid specifically for these placements, AI Max-enabled campaigns are eligible for this inventory.
Google's headline numbers are well-known: 14% average conversion improvement, 27% lift for campaigns heavily using exact and phrase match. These come from Google's internal data and represent aggregate averages across a large sample. They are directionally useful but hide enormous variance.
Independent testing across hundreds of accounts tells a more nuanced story. Approximately 84% of advertisers who have tested AI Max report neutral or negative results. This is not because AI Max does not work. It is because most accounts are not properly prepared for it. The 16% who see positive results share common characteristics: strong conversion tracking with Enhanced Conversions enabled, comprehensive negative keyword lists, quality landing pages across their site (not just one or two optimized pages), adequate conversion volume (100+ monthly conversions), and established success with broad match keywords before enabling AI Max.
The accounts that fail with AI Max almost always share a different set of characteristics: thin negative keyword lists that allow AI Max's expanded matching to generate waste, poor conversion tracking that gives the AI bad signals, limited landing page quality (one good page surrounded by mediocre ones, which hurts final URL expansion), insufficient budget (if you are already budget-constrained, AI Max diverts spend from your proven keywords to experimental queries), and no prior broad match testing (if broad match has not worked for you, AI Max, which expands matching even further, is unlikely to improve things).
Based on data from across the industry through early 2026, here is what to expect. For well-prepared accounts meeting all prerequisites, a 10-20% conversion improvement at similar CPA is realistic. Achieving a 10% improvement puts you ahead of the vast majority of AI Max testers. For accounts with some readiness gaps (decent but not comprehensive negative keywords, adequate but not exceptional conversion tracking), expect roughly flat performance with some incremental reach. For accounts that are not ready (thin negatives, manual bidding, limited conversion data), expect higher CPAs and wasted spend. Do not enable AI Max on these accounts until the foundation is solid.
AI Max amplifies whatever is already happening in your account. If your account is well-optimized, AI Max amplifies good results. If your account has structural issues, AI Max amplifies those too. Meeting these prerequisites is not optional. It is the difference between the 16% who succeed and the 84% who do not.
This is non-negotiable. AI Max needs accurate conversion signals to learn which queries and ads drive results. If your conversion tracking overcounts (tracking page views as conversions) or undercounts (missing offline conversions, not tracking phone calls), AI Max will optimize toward the wrong outcomes. Enhanced Conversions should be enabled to recover conversion data lost to cookie restrictions and cross-device paths.
AI Max only works effectively with automated bidding strategies: Target CPA, Target ROAS, Maximize Conversions, or Maximize Conversion Value. Manual CPC bidding does not provide the signals AI Max needs. Industry testing shows significantly more stable results with Target CPA and Target ROAS than with Maximize strategies, because Target strategies constrain the AI to efficiency thresholds while Maximize strategies allow unlimited CPA/ROAS variance.
Google technically allows AI Max with any conversion volume, but practical testing shows that accounts with fewer than 30 monthly conversions produce highly erratic AI Max results. For reliable performance, aim for 100+ monthly conversions. Below 30, the AI simply does not have enough data to learn effectively.
If your campaigns are already losing impressions due to budget constraints, enabling AI Max usually makes things worse. AI Max diverts some of your budget from proven keywords to experimental query matching. If you do not have budget headroom, this diversion reduces spend on what works without generating enough data on the experimental queries to determine what is working and what is not. Google recommends a daily budget of at least 15 times your target CPA.
This is where most AI Max implementations fail. AI Max expands matching aggressively, which means it will serve your ads for queries your keyword list never anticipated. Without a comprehensive negative keyword strategy, these expanded matches generate significant waste. Build out your negative keyword lists before enabling AI Max, not after. At minimum, you should have account-level negatives for job queries, free/bargain intent, explicit content, and irrelevant industries. Campaign-level negatives should cover industry-specific irrelevant terms. See our complete negative keyword guide for a 500+ keyword starter list organized by industry.
AI Max treats all keywords as effective broad match and then expands further. If you have not successfully used broad match keywords in your account, AI Max will almost certainly underperform. Test broad match with Smart Bidding on a few ad groups first. If broad match produces acceptable results, your account is ready for AI Max. If broad match generates excessive waste even with negative keywords, fix the broad match problem before adding AI Max on top of it.
Meeting the prerequisites gets you to the starting line. These best practices determine how well you actually perform.
Never enable AI Max across your entire account simultaneously. Use the built-in one-click experiment feature to run a 50/50 split test on your highest-performing campaign. Run the experiment for a minimum of 4 weeks, ideally 6 weeks, to allow the learning period to complete and reach statistical significance. Only scale to additional campaigns after the experiment shows clear positive results.
AI Max will surface query patterns you have never seen before. Some will be valuable discoveries. Others will be irrelevant waste. During the first 2-4 weeks after enabling AI Max, check your search terms report daily to identify new patterns that need negative keywords. This front-loaded monitoring investment pays for itself by catching waste before it accumulates.
As text guidelines roll out through Q1 2026, configure them immediately. Exclude terms that do not align with your brand positioning. Provide clear guidance about your brand voice, value propositions, and messaging priorities. Do not rely on AI Max to infer your brand identity from your landing pages alone. Be explicit.
Final URL expansion is enabled by default with AI Max, and it can be either a major advantage or a significant problem. If your entire site is well-optimized with strong landing pages across all key sections, final URL expansion helps by matching users to the most relevant page. If your site has inconsistent quality (some great pages, some outdated pages, some pages with no clear CTA), final URL expansion will send some traffic to underperforming pages, which hurts conversion rates.
Review which pages AI Max sends traffic to through the landing page report. Use URL exclusions to block pages that should not receive ad traffic (blog posts, about pages, career pages, policy pages). Use URL inclusions at the ad group level to constrain expansion to specific sections of your site.
The combination of AI Max and Smart Bidding Exploration is powerful but aggressive. Both features expand what your campaigns can reach: AI Max expands queries, Smart Bidding Exploration expands bidding flexibility. Together, they cast a much wider net. Only combine these if your conversion volume is high (200+ monthly conversions), your negative keyword infrastructure is robust, and you can tolerate a short-term CPA increase while the system learns. For most accounts, enable AI Max first, stabilize performance, and then layer Smart Bidding Exploration on top.
A common misconception is that AI Max makes keywords irrelevant. It does not. Your keywords still serve as critical signals that guide AI Max's matching. Well-organized, intent-specific keywords give the AI a better foundation to expand from. Poorly organized, overly broad keywords give the AI ambiguous signals that lead to worse expanded matches. Continue managing your keyword structure even with AI Max enabled.
The three-way choice between AI Max Search, Performance Max, and standard Search is the defining strategic question for Google Ads in 2026.
You need search-specific visibility with the ability to see exactly which queries trigger your ads. You want to expand reach beyond your keyword list while maintaining Search campaign transparency. Your business is B2B, lead generation, or services where query-level visibility and messaging precision matter. You have the conversion tracking, negative keyword infrastructure, and budget to support AI Max's expanded matching. You are testing the waters with automation while retaining more control than Performance Max offers.
You need cross-channel reach beyond Search (Display, YouTube, Gmail, Shopping, Discover, Maps). You are in ecommerce with a strong product feed and visual assets. You want full-funnel coverage from awareness through conversion in a single campaign. You prioritize total conversion volume over query-level transparency.
You need absolute control over which queries trigger your ads. You are in a regulated industry where every piece of ad copy must meet compliance requirements. Your budget is constrained and you cannot afford to test expanded matching. Your conversion volume is below the threshold for AI Max to learn effectively (under 30 monthly conversions).
Google now promotes the "Power Pack" strategy: Performance Max for broad multi-channel reach (60-70% of budget for ecommerce, 30-40% for B2B), AI Max for Search for high-intent search coverage with transparency (30-40% of budget), and Demand Gen for top-of-funnel awareness. This structure gives each campaign type a distinct role with minimal overlap when properly configured with negative keywords preventing cannibalization between campaign types.
AI Max adds a powerful layer of Google's AI to your Search campaigns. But it also adds complexity. More query patterns to monitor. More ad copy variants to evaluate. More landing page combinations to track. More negative keyword decisions to make. More potential for waste if the expanded matching hits irrelevant territory.
For advertisers managing campaigns manually or with semi-autonomous tools, AI Max creates more work, not less. The expanded matching requires more frequent search term reviews. The dynamic ad copy needs more monitoring for brand consistency. The final URL expansion needs ongoing page-level performance analysis. The interaction with Performance Max needs careful management to prevent cannibalization.
This is where autonomous management fundamentally changes the equation. groas integrates with AI Max the way autonomous systems are meant to: by working with Google's AI rather than fighting it, while independently ensuring that Google's expanded matching serves your business interests rather than Google's revenue interests.
Specifically, groas monitors the query patterns AI Max generates in real time, adding negative keywords within hours when irrelevant patterns emerge (versus the days or weeks it takes with manual or semi-autonomous management). It evaluates which AI Max-matched queries actually convert at acceptable CPAs and adjusts campaign signals to reinforce valuable patterns. It monitors the interaction between AI Max-enabled Search campaigns and Performance Max campaigns, managing negative keywords and search themes across both to prevent cannibalization. It assesses landing page performance for URLs served through final URL expansion, identifying pages that receive AI Max traffic but underperform on conversion rate. And it does all of this continuously across all twelve optimization levers simultaneously, which is the critical difference.
AI Max is a good feature. It genuinely finds queries and conversions that traditional keyword management misses. But like any powerful tool, it requires active, continuous management to deliver its potential. The advertisers who get the 14%+ improvement are not the ones who enable AI Max and walk away. They are the ones whose accounts are continuously optimized around the expanded matching that AI Max provides. groas makes that continuous optimization automatic.
Based on Google's strategic direction, product announcements, and the pattern of updates through 2025, here is what is likely coming.
Text guidelines expanding to all advertisers globally, providing full brand control over AI-generated copy in both AI Max and Performance Max. Enhanced experiment reporting with more granular breakdowns of AI Max performance by query source, creative variant, and landing page. Google is likely to push AI Max as the default recommendation for all new Search campaigns.
Dynamic Search Ads may receive a formal deprecation announcement with a 12-18 month migration timeline. The overlap between DSA, AI Max, and Performance Max's keywordless matching makes DSA increasingly redundant. Separate bidding controls for AI Max versus traditional search placements may emerge, giving advertisers more control over how aggressively AI Max bids on expanded queries. More granular controls for AI Overviews and AI Mode ad placements, which are currently served automatically without advertiser control over these specific surfaces.
Google's trajectory is clear: Search campaigns are moving toward AI-driven intent matching rather than manual keyword management. AI Max is the transitional technology that maintains keyword infrastructure while introducing AI-powered expansion. Over time, expect keywords to become more like intent signals than exact targeting mechanisms, with AI Max (or its successors) handling the actual query matching. The advertisers who thrive will be those who learn to guide the AI effectively through signals (search themes, negative keywords, audience data, creative direction) rather than those who try to maintain manual control over every query match.
AI Max is an optional suite of AI-powered features that you enable on existing Search campaigns. It includes three components: search term matching (expanding your reach beyond your keyword list using broad match and keywordless technology), asset optimization (dynamically generating ad copy using Google's Gemini model and selecting optimal landing pages), and Locations of Interest (targeting based on geographic intent). It is not a new campaign type. It is an enhancement layer that gives your Search campaigns some of Performance Max's intelligence while retaining Search's transparency and control.
For accounts that meet the prerequisites (accurate conversion tracking, automated bidding, 100+ monthly conversions, comprehensive negative keywords, proven broad match success, and adequate budget), AI Max is worth testing. A realistic expectation for well-prepared accounts is a 10-20% conversion improvement at similar CPA. For accounts that are not ready, enabling AI Max typically increases waste without improving results. Use the built-in experiment feature to test on a single campaign before committing.
The primary reasons are inadequate negative keyword lists (AI Max expands matching aggressively, generating waste without proper exclusions), poor conversion tracking (giving the AI bad signals to learn from), budget constraints (AI Max diverts spend from proven keywords to experimental queries), no prior broad match success (AI Max amplifies matching even further than broad match), and insufficient conversion volume (below 30 monthly conversions, the AI cannot learn effectively). The 16% who succeed share well-prepared accounts across all these dimensions.
AI Max is an enhancement to Search campaigns. Performance Max is a standalone campaign type serving ads across all Google channels. AI Max gives you full search terms transparency, keyword-level control, and familiar Search campaign reporting. Performance Max offers broader reach across Display, YouTube, Gmail, Shopping, Discover, Maps, and Waze, but with less transparency and control. AI Max is generally better for B2B, lead generation, and services. Performance Max is generally better for ecommerce with large product catalogs and strong visual assets.
Very likely. The overlap between DSA, AI Max, and Performance Max's keywordless search matching makes DSA increasingly redundant. Industry experts predict a formal deprecation announcement in Q2 2026 with a 12-18 month migration period. If you rely on DSA campaigns, begin testing AI Max as a replacement now.
Yes, increasingly so. Text guidelines, rolling out through Q1 2026, allow you to exclude specific terms the AI should never use and provide natural language guidance about your brand voice and messaging priorities. You can also pin certain headlines and descriptions in your responsive search ads to ensure key messages always appear. However, AI Max can override pinned assets if final URL expansion is enabled and the AI identifies a more relevant landing page.
AI Max requires more aggressive negative keyword management than standard Search campaigns because it matches queries more broadly. At minimum, implement account-level negatives for universal categories (job queries, free/bargain intent, explicit content). Add campaign-level negatives for industry-specific irrelevant terms. During the first 2-4 weeks after enabling AI Max, review search terms daily to catch new irrelevant patterns the expanded matching surfaces. Ongoing, review at least weekly. This is one of the primary reasons autonomous management outperforms manual management with AI Max. groas catches and blocks irrelevant patterns within hours, versus the days or weeks it takes with periodic human review.
AI Max requires automated bidding (Target CPA, Target ROAS, Maximize Conversions, or Maximize Conversion Value). Manual CPC does not provide the signals AI Max needs. Independent testing shows significantly more stable results with Target CPA and Target ROAS than with Maximize strategies, because Target strategies constrain the AI to efficiency thresholds while Maximize strategies can lead to runaway CPAs.
This is a critical structural question. AI Max-enabled Search campaigns and Performance Max both compete for search inventory. Google generally gives priority to Search campaigns with exact match keywords, but AI Max's expanded matching creates overlap with PMax's search inventory. To prevent cannibalization, add your core Search keywords as negatives in Performance Max, use search themes in PMax to guide it toward queries your Search campaigns do not cover, and monitor search terms reports in both campaign types for overlap.
groas integrates with AI Max by managing the continuous optimization that AI Max requires to perform well. This includes real-time negative keyword management as AI Max surfaces new query patterns, cross-campaign cannibalization prevention between AI Max Search and Performance Max, landing page performance monitoring for URLs served through final URL expansion, ad copy quality assessment for AI-generated variants, and coordinated optimization across all twelve campaign levers simultaneously. groas works with Google's AI (including AI Max, Smart Bidding, and Performance Max) rather than against it, while ensuring that the expanded automation serves your business goals rather than Google's revenue targets.
