Google AI Max for search campaigns: Fully automated advertising – 14% more conversions or a costly loss of control for advertisers?

Beware of hidden costs: Why you should never activate Google AI Max without these 3 safeguards

Performance Max vs. AI Max: Why Google's new AI is changing search campaigns forever

With "AI Max," Google has launched what is arguably its most ambitious automation project for search campaigns. The promise sounds enticing: an average of 14 percent more conversions without requiring any structural changes to the account. A single click is all it takes, and the AI ​​takes over. But behind this gleaming facade lies a technological paradigm shift that is tearing down fundamental pillars of traditional search engine marketing. Flipping the switch means relinquishing precise control over search queries, ad copy, and landing pages to a black box. While Google's own case studies celebrate impressive successes, independent analyses paint a much more nuanced picture, ranging from significant performance leaps to massive budget losses and legal compliance risks. However, with the mandatory migration of Dynamic Search Ads (DSA) at the beginning of 2027, there will be no way around AI Max. This article sheds light on how the new system really works, which figures Google prefers to keep quiet about, and which specific strategies advertisers can use to tame AI in order to effectively protect their brand and budget.

When the algorithm takes over – what advertisers really need to know about Google's most powerful automation tool

Google's next step towards fully automated advertising

In May 2025, Google launched AI Max for search campaigns, a product that can be described as the most ambitious automation project in the history of Google Ads. It's not a new campaign type, but rather an optimization layer that can be integrated into existing search campaigns with a single click, fundamentally changing how they function. The announcement was accompanied by a figure that caused a stir in the marketing world: advertisers who activate AI Max would achieve an average of 14 percent more conversions or conversion value with comparable CPA or ROAS. For campaigns still primarily based on exact and phrase-match keywords, the typical uplift is even higher, reaching 27 percent.

Google presented a message that sounds deceptively clear at first glance: more performance without structural changes. But, as is so often the case with technological leaps in digital marketing, the devil is in the details. Activating AI Max means relinquishing fundamental control mechanisms that have so far been considered inviolable cornerstones of professional search campaigns: precise control over matched search queries, displayed ad copy, and the landing pages to which users are directed. The central question that preoccupies experienced advertisers is therefore not: Does AI Max work? But rather: Does it work for me, in my way, and under my conditions – without me losing strategic control?

Three levers, one black box: The technical architecture of the system

AI Max combines three closely intertwined functions under a common technical architecture. The first and most far-reaching element is Search Term Matching, which is based on a combination of broad match and keywordless technology. The system analyzes existing keywords, creative assets, and URLs, and learns from this input to deliver ads for new, previously unaddressed search queries that it deems relevant. At its core, this principle is an enhanced form of broad match, but it goes further by delivering ads even when no matching keyword variant exists in the account.

The second element is text adaptation, formerly known as "Automatically Created Assets," which is now a mandatory component of AI Max once Final URL Expansion is enabled. The system dynamically generates ad titles and descriptions from landing page content, existing ad copy, and keyword information. It uses Google's generative AI to create text that matches the search query, not necessarily the advertiser's editorial guidelines. The third element, Final URL Expansion, automatically redirects users to the most relevant subpage of the website, as determined by the algorithm, regardless of the URL originally used in the ad.

What technically connects these three components is the principle of predictive intent recognition. Google states that it no longer simply reacts to past search queries, but predicts what users might search for next and delivers ads in moments and contexts that were previously inaccessible to paid search advertising. This sounds efficient, but at the same time represents a departure from the classic deterministic logic of keyword-based search engine marketing: away from the equation "keyword equals ad equals landing page," towards a probabilistic modeling of user behavior, in which the algorithm makes situational decisions that no human campaign manager has defined or approved in advance.

What the data really shows – and what Google is hiding

Google's own performance metric of 14 percent more conversions at the same CPA sounds convincing. However, a closer look reveals methodological limitations that should raise red flags for any serious performance marketer. First, the figure is based on internal Google data from 2025 and refers exclusively to non-retail advertisers. E-commerce companies, one of the largest and most important advertiser groups in digital marketing, are explicitly excluded from this benchmark. Google mentions this limitation in a footnote, not in the headline.

The first independent large-scale study, published in March 2026 and based on the analysis of more than 250 Google Ads campaigns, paints a much more nuanced picture. Median revenue did indeed increase by 13 percent, which comes close to Google's promise. At the same time, however, the median CPA rose by 16 percent. The ROAS fluctuated between plus 42 and minus 35 percent, indicating extreme heterogeneity in the results. Mike Ryan of Smarter Ecommerce, who conducted the analysis, succinctly summarized the findings: Activating AI Max is, in many cases, like a coin toss – you might get a lift, but the efficiency usually doesn't keep pace.

A separate, independent analysis from November 2025, encompassing over 250 campaigns, revealed that AI Max achieved up to 35 percent lower ROAS compared to traditional match types. These figures stand in stark contrast to official Google communications and suggest that the system is by no means automatically the right choice in industries and contexts where efficiency trumps volume. The variance in results is the real issue: AI Max can perform brilliantly or cause significant budget losses, and which scenario will unfold is virtually impossible for individual advertisers to predict.

The promise problem: Why Google's own numbers call for caution

Google's communication strategy surrounding AI Max follows a pattern familiar from the history of digital advertising platforms: performance data is measured under optimal conditions, cited from case studies, and then communicated as an average value without specifying the conditions under which it applies. Case studies like those from L'Oréal, which reported a doubled conversion rate with a 31 percent lower cost-per-conversion, or from MyConnect Australia, which recorded 16 percent more leads with a 13 percent lower CPA, are real, but also selective.

What's missing is a representative disclosure of the overall results distribution. What percentage of advertisers actually experienced an improvement, and how many a decline? The answer to this question is not found in Google's official communications. Independent analyses partially fill this gap, but they too are not without limitations, as the analyzed campaigns often come from a specific mix of agency clients. Overall, the data reveals that AI Max is not a universal performance upgrade, but rather a context-dependent tool with significant upside potential and equally significant downside risk. The systematic omission of the retail exclusion clause in the main communications is particularly problematic, as e-commerce companies are among the largest Google Ads investors.

Furthermore, in April 2026, Google published an updated performance metric that initially appears confusing: The full suite of AI Max, combining search term matching, text optimization, and final URL expansion, yields an average of 7 percent more conversions than search term matching alone. This figure sounds lower than the original 14 percent because it uses a different reference point. It measures the incremental uplift from the creative and landing page components compared to the targeting uplift alone—a distinction easily overlooked in the daily hustle and bustle, but fundamental for the strategic evaluation of the system.

The illusion of control: Where real control possibilities end

From the outset, Google accompanied AI Max with a promise designed to allay advertisers' skepticism: increased performance with maintained control. Indeed, the system offers a range of control mechanisms that go beyond what Performance Max allows. Brand Controls enable the inclusion or exclusion of specific brands, preventing ads from appearing alongside unwanted brand terms. Locations of Interest allow for location-based targeting at the ad group level. URL inclusion and exclusion rules give advertisers the ability to prioritize or block specific landing pages. Negative keywords are respected in AI Max.

However, the crucial difference to traditional manual campaign management lies in the fact that all these control mechanisms are reactive. Ads can be removed after they've started running, but not approved in advance. URLs can be excluded after the system has misidentified them, but you don't define in advance which pages should be targeted. Negative keywords can be added after a costly mismatch has been identified, but you can't proactively prevent them. Text generation, in particular, is problematic from a compliance perspective: The system can generate dozens of new ad variations daily that no human reviewer has seen before they are launched.

A specific case that caused a stir in professional circles illustrates the structural problem: A British financial services brand discovered that AI Max contained automatically generated assets that implicitly suggested the service did not require a credit check—a legally mandated disclosure under British financial law. The legal implications of such an automatically generated, unlawful statement are borne not by Google, but by the advertiser. This is not a hypothetical scenario, but a documented incident demonstrating the significant gap between Google's "We use your approved assets" claim and the reality of AI-generated text.

Who benefits the most – and who should be especially careful

Based on the available data and practical experience, a differentiated profile of AI-Max winners and losers can be drawn. The system performs particularly well for advertisers with large inventories or service portfolios, as keyword-free matching can fill genuine gaps in coverage. Companies with small teams that benefit from automation without having to prioritize granular control also belong to the winning group. E-commerce brands with a strong performance focus in the non-branded sector, who are prepared to justify the higher CPAs through increased sales, can also benefit.

In contrast, particular caution is advised in highly regulated sectors such as healthcare, finance, and legal, as automated text generation can easily veer into compliance-critical territory. Brands with clear tone-of-voice guidelines and restrictive brand policies risk AI Max generating messages that don't align with the brand's tone of voice. Advertisers with very tight budgets and high CPA sensitivity are also ill-advised to activate AI Max without prior A/B testing, as the potential CPA increase of 16 percent or more can jeopardize overall campaign efficiency in the short term.

The situation is particularly interesting for advertisers currently using Dynamic Search Ads (DSA). In April 2026, Google announced that it would discontinue DSA as a standalone format at the beginning of 2027 and automatically migrate all affected campaigns to AI Max. Following significant advertiser resistance, the original September 2026 deadline for DSA campaigns was postponed to February 2027. Those who passively wait for the migration risk Google choosing default settings optimized for maximum reach, potentially leading to initial budget inefficiencies. Proactive, self-directed migration gives advertisers the opportunity to configure settings, exclusions, and URL rules before the learning phase begins.

Xpert.Digital is a data-driven B2B industry hub led by Konrad Wolfenstein . The company acts as an external, quasi-in-house solution for industrial partners, closing operational gaps in marketing, content, and sales – without requiring additional resources on the client side.

More information here:

• The quasi-in-house solution: How Xpert.Digital closes operational gaps in B2B marketing and sales – Smart Content-Driven Business

AI Max in system comparison: How it differs from Performance Max

A key misunderstanding in the discussion surrounding AI Max is its conflation with Performance Max. Both systems rely on AI-powered automation, but they follow fundamentally different philosophies. Performance Max is a cross-channel system that allocates budget across search, display, YouTube, Discovery, Gmail, and Maps. The advertiser provides assets and objectives; the algorithm decides everything else. Keyword control is nonexistent, and granular query-level reporting is virtually impossible.

AI Max, on the other hand, remains in the search channel and combines AI automation with a higher level of transparency and control than Performance Max. Negative keywords work, search term reports are available, and URL control is possible at the campaign and ad group levels. In a study of 24,702 campaigns, Search performed almost twice as well as Performance Max in terms of conversion rate—an argument that AI Max is systematically preferable for high-intent B2B and high-consideration industries.

The relevant conclusion for strategically minded advertisers is this: AI Max is not a step towards Performance Max, but rather an AI extension of the traditional search channel that complements, not replaces, keywords. Google itself emphasizes that keywords remain central to campaign structure because they provide the intent signals upon which the algorithm is built. Those who dismantle their keyword structure in the belief that AI Max will completely take over this function risk degrading the data foundation on which the system is trained.

The new control instrument: AI Brief and Text Guidelines

One of the most important recent developments surrounding AI Max is the AI ​​Brief feature, introduced in April 2026 and based on Google's Gemini model. AI Brief allows advertisers to tell the AI ​​system in natural language what ads should and shouldn't communicate, who they should target, and which matching criteria apply. Specifically, messaging guidelines such as "Never mention prices," matching guidelines such as "Prioritize searches for healthy staples," and audience guidelines such as "For health-conscious users: Highlight our clean-label products" can be defined.

AI Brief generates previews of assets and search queries, allowing advertisers to provide feedback and make adjustments before the campaign launches. This represents a significant conceptual improvement over the previous system, where AI-generated text only became visible after the ad was sent. AI Brief is complemented by Text Guidelines, which allow advertisers to exclude up to 25 specific terms from text generation and define up to 40 content restrictions. For highly regulated industries, the Text Disclaimers feature has also been introduced, ensuring that legally required disclosures appear in ads even when Final URL Expansion is enabled.

These developments show that Google is responding to advertiser feedback and gradually addressing the control deficit of the initial phase. At the same time, they make it clear that the system carries significant risks without conscious configuration of these guidelines. AI Brief and Text Guidelines are not automatically active and require proactive maintenance. Advertisers who activate AI Max and ignore these features effectively have no control over the AI-generated content of their ads.

The right activation strategy: Test logic before rollout mentality

The biggest strategic mistake when implementing AI Max is indiscriminately activating it on all campaigns simultaneously. The correct approach is based on a clear testing logic that utilizes Google's built-in Experiments feature. This feature, found under the "Experiments" menu item in the Campaigns section, allows for a 50/50 split test within an existing campaign without having to create a copy. It divides traffic and budget within the running campaign, with one half running with AI Max enabled and the other without.

Several factors must be considered for a meaningful test. First, the daily budget should be at least €50, as Google itself advises against using AI Max with campaigns that have lower budgets. Second, the campaign should provide sufficient conversion data to achieve statistical significance, which requires a minimum duration of four to six weeks. Third, a clear baseline report should be exported before the test, including conversion data, search term reports, and landing page metrics to allow for before-and-after comparisons.

Alongside the testing phase, three configuration measures are essential: First, defining a comprehensive negative keyword list that includes all irrelevant categories and known problematic terms. Second, setting up brand controls that either move your own brand terms to separate brand campaigns or manage them via brand inclusions and exclusions. Third, configuring URL exclusion rules for pages unsuitable as landing pages, such as imprint pages, career subpages, or purely informational article pages. These three measures together form the safety net that makes AI-Max activation truly responsible.

Fire and budget protection as a strategic architectural task

Protecting brand and budget in an AI Max environment isn't a matter of individual settings, but rather an architectural design challenge affecting the entire campaign structure. The most important measure is the strict separation of brand and non-brand campaigns. Brand campaigns should generally not activate AI Max, as the risk of wasting budget through keyword cannibalization and misallocation in a highly competitive auction is real. The keywordless matching technology can lead to brand campaigns spending budgets on search queries that would already be well covered by organic results.

Structured URL governance is the second key pillar of budget protection. Final URL expansion is a powerful tool, but only if the entire website meets the quality requirements for paid traffic. Pages with weak conversion infrastructure, missing call-to-action elements, or inadequate mobile optimization should be actively excluded from URL expansion. The AI-Max system selects landing pages based on relevance from a search engine perspective, not conversion probability—which is why manually curating the allowed URL base is non-negotiable.

The third safeguard is weekly search term monitoring with clearly defined escalation thresholds. Search queries that generate spend above a defined threshold without delivering conversions must be immediately added as negative keywords. The AI-Max system learns from conversion signals, and a lack of early-stage restrictions can lead to the development of inefficient patterns that are harder to correct than with traditional keyword campaigns. Structured weekly reporting that groups costs, impressions, and conversions by search term category is the necessary foundation for data-driven optimization.

The DSA migration as a turning point: What needs to be done strategically now

The upcoming migration from Dynamic Search Ads to AI Max is not just a routine technical task, but marks a strategic turning point for any advertiser who uses DSA as a cornerstone of their search strategy. DSA will automatically switch to AI Max in February 2027, and the default settings Google uses for the auto-migration are geared towards maximum reach, not maximum efficiency.

Advertisers who proactively manage the migration process have several months to establish a clean data foundation. This includes exporting historical DSA reports as a performance baseline, mapping existing DSA targeting rules to the corresponding AI Max URL inclusion and exclusion rules, and thoroughly reviewing and updating negative keyword lists. Google provides upgrade tools that transfer historical settings and data to new default ad groups to ensure a smooth transition.

It is particularly important to understand that AI Max and DSA are conceptually different systems: While DSA deterministically analyzes landing pages and generates headlines from them, AI Max works predictively and generatively by using real-time intent signals and dynamically creating ad content that fits the user context, not just the underlying website. This conceptual difference means that a DSA campaign that performs exceptionally well will not automatically perform as well after migration until the AI ​​system has collected enough conversion data to calibrate itself. This learning phase must be planned for and secured with sufficient budget and time buffers.

The bigger picture: What Google's automation offensive means for the industry

AI Max shouldn't be viewed in isolation, but rather as part of Google's systematic strategy to gradually replace manual control with AI-powered automation. This progression extends from Smart Bidding through Responsive Search Ads and Performance Max to AI Max and AI Brief: at each step, a portion of manual control is traded for promised performance gains. The pattern is clear, and it would be naive to consider AI Max the final stage in this evolution. Anyone adopting AI Max today is preparing for an advertising ecosystem where the ability to configure and control AI systems becomes more important than the ability to manage keywords.

This has far-reaching implications for skills development in digital marketing. Traditional SEM expertise—keyword research, match type strategy, manual bid optimization—is gradually losing importance. Instead, the ability to guide AI systems with high-quality input is gaining in significance: precisely defined conversion goals, well-structured website architectures with clear URL hierarchies, complete and maintained asset libraries, and disciplined exclusion governance. Advertisers who continue to try to run search campaigns as they did in 2015 will increasingly fall behind.

At the same time, the industry should remain vigilant regarding a dynamic often observed in the history of commercial platforms: the more advertisers rely on the proprietary automation of a single provider, the less bargaining power they have and the less control they have over the efficiency of their media spend. AI Max is a powerful tool, but it is Google's tool, and its performance parameters are not defined independently, but by a company whose core business model is based on maximizing advertising spend. Therefore, the critical, data-driven evaluation of the system, as already conducted by independent studies, is not mere nitpicking, but a professional obligation.

☑️ Our business language is English or German

☑️ NEW: Correspondence in your native language!

 

I and my team are happy to be available to you as your personal advisor.

You can contact me by filling out the contact form here wolfenstein@xpert.digital:or simply call me at +49 7348 4088 965. My email address is

I'm looking forward to our joint project.

 

 

☑️ SME support in strategy, consulting, planning and implementation

☑️ Creation or realignment of the digital strategy and digitization

☑️ Expansion and optimization of international sales processes

☑️ Global & Digital B2B trading platforms

☑️ Pioneer Business Development / Marketing / PR / Trade Fairs

In industrial B2B, sustainable business relationships rarely emerge overnight. They develop step by step – through visibility, professional relevance, recurring touchpoints, and growing trust. Xpert.Digital's 4-stage model addresses precisely this: It offers a structured path that begins with a manageable entry point and can evolve into deeper collaboration in business development if needed.

Instead of relying on loud marketing promises, this model puts the relationship at the forefront. Companies start with clearly defined, easily calculable measures and then decide, based on their own experience, how far they want to expand the collaboration. A key factor for this undisturbed trust-building process: The platform completely avoids annoying advertising ads, so the editorial focus remains solely on the companies' expertise.

More information here:

• From visibility to trust: Your scalable path with Xpert.Digital