Marketers Shift Budgets to AI Search as Generative Engine Optimization Overtakes Traditional SEO

Marketing Budgets Shift Toward AI Search Optimization

PMG, one of the world's largest independent marketing agencies, now advises clients to pilot generative engine optimization at 1.5 to 2 times their existing search budget, according to Matt Allfrey, its head of SEO EMEA2

2

. A June 2026 Semrush study of nearly 500 marketing professionals found that more respondents plan to invest in AI search optimization at 38% compared to traditional SEO at 36%, marking a quiet crossover moment for an industry that treated generative engine optimization as a curiosity just two years ago2

2

. In the same survey, 85% of marketers said AI has already changed how they approach search2

2

.

This shift reflects a fundamental restructuring of how people discover information online. AI-powered search engines now process questions through a retrieve, rerank, and generate pipeline that leaves little room for weak or poorly structured content4

4

. Google AI Overviews now reach more than 2.5 billion monthly users, while ChatGPT serves around 1 billion weekly users4

4

. AI visibility now depends on retrieval quality instead of search ranking alone, as content that never reaches the first retrieval stage cannot appear in the final response, no matter how detailed it looks4

4

.

Model Upgrades Replace Algorithm Updates

On January 27, 2026, the AI visibility of thousands of businesses changed overnight when Google quietly swapped a new model, Gemini 3, into AI Overviews and AI Mode1

1

. The number of cited sources per AI answer jumped by roughly a third, freshness suddenly carried more weight, and entity-rich websites gained share at the expense of thinner ones1

1

. ChatGPT, running on an entirely separate pipeline, was completely unaffected1

1

.

Model upgrades have become the AI-era equivalent of algorithm updates, arriving unannounced, undocumented, and on multiple platforms at once1

1

. Modern AI search systems fan out a single question into many parallel sub-queries, often eight to twelve, and in ChatGPT's case up to twenty, before retrieving sources for each, verifying claims, and synthesizing an answer1

1

. Research already shows that only a quarter to a third of AI citations come from pages ranking in the traditional top ten1

1

. With Gemini 3.5 Pro arriving and further flagship releases expected from every major lab before year's end, businesses should plan for several more of these invisible resets1

1

.

Three Diverging AI Search Ecosystems

AI search is no longer one thing as the three dominant assistants diverge into fundamentally different strategies1

1

. ChatGPT is doubling down on personalization and monetization with its memory features maturing rapidly and its advertising pilot expanding internationally to the UK, Mexico, Brazil, Japan, and South Korea1

1

. Gemini is fusing Google's retrieval and trust infrastructure with in-chat commerce, letting users complete purchases without ever leaving the conversation1

1

. Claude has positioned itself as the ad-free option focused on professional and agent-driven work1

1

.

One large-scale citation study found that only 11% of domains are cited by both ChatGPT and Perplexity, and brand recommendations can differ by 40 to 60% across platforms for identical queries1

1

. Each engine is its own ecosystem with its own biases and blind spots1

1

. A dropshipping platform that was being recommended heavily by Gemini as a top option had vanished entirely from certain ChatGPT recommendations despite having been ChatGPT's number-one pick for the same prompts just two months earlier1

1

. When asked why, ChatGPT explained that the platform was not known for working well with Shopify, even though the prompt had never mentioned Shopify at all1

1

.

Google Publishes Official GEO Guidelines

In mid-May, Google released its first official guide to optimizing for generative AI features, a document that addresses in plain language what influences AI visibility in AI Overviews and AI Mode2

2

. The guide's central message is that optimizing for generative AI search is, from Google's perspective, still SEO2

2

. The guide explicitly debunks several practices sold aggressively under the generative engine optimization banner over the past two years, stating there is no special schema markup that unlocks AI visibility and chasing inauthentic mentions of your brand across the web is far less effective than it appears2

2

.

In early June, Google updated its long-standing guidance on hiring SEO help, adding a new document on evaluating third-party tools and services and explicitly naming GEO and AEO as service categories for the first time2

2

. According to reporting by Nieman Lab, referral traffic from search engines has fallen by roughly 60% for small publishers and 47% for medium-sized ones over the past two years2

2

. These numbers highlight a clear shift from page rankings toward AI citations and brand mentions as AI-generated answers become the primary product4

4

.

Winner-Takes-All Dynamics in Citation-Based Search

AI visibility operates as a winner-takes-all game far more than classic Google search2

2

. When an AI answer engine recommends two or three options in a conversational answer, most users simply accept them without scrolling through alternatives, opening ten tabs, or venturing to page two2

2

. The gap between being cited and not being cited is no longer a difference in degree but a fundamental difference in market access2

2

.

AI search tools, answer engines, and citation-heavy summaries are reshaping how buyers evaluate local services, exporters, and professional firms5

5

. Visibility no longer starts on a homepage but in the answers people read before they ever click5

5

. Zero-click searches have increased as AI assistants answer a large share of online questions before visitors open traditional search results4

4

.

Regulated Industries Gain Advantage Through Verifiable Credentials

Regulated businesses in healthcare and law now hold an unfair advantage in AI search3

3

. The caution these firms are forced to carry, including named experts, verified claims, and outside validation, is the exact signal AI answer engines look for when they decide whom to cite3

3

. When the cost of being wrong is high, the engine hedges toward sources that lower its risk, favoring accuracy, verifiable credentials, and outside validation3

3

.

This is sharpest in what search systems classify as YMYL, your money or your life, the health, legal, and financial queries where a bad answer does real damage3

3

. In law, AI-driven search tends to name firms with peer recognition, presence in the directories the profession already trusts, and consistent public information across the web3

3

. In healthcare, it favors authoritative content with named clinicians, visible medical review, real credentials, and information that is current and accurate3

3

.

What Works in the New Retrieval Process

The original GEO study found that content with statistics, quotations, citations, and reliable references achieved visibility gains of up to 40% within controlled experiments4

4

. A larger review of 45 GEO studies, published in July 2026, reached another important conclusion: content structure, topical relevance, and strong context placement deliver more reliable results than generic optimization tricks4

4

.

Clear headings, direct answers, well-organized sections, trusted statistics, and accurate citations help retrieval systems understand each page with greater confidence4

4

. Independent publications and respected news outlets strengthen authority as many AI systems trust third-party validation more than promotional claims from company websites4

4

. Strong entity descriptions, structured information, and fresh updates increase the chance of selection during retrieval and citation stages4

4

. Earned media and expert commentary now function as distribution because when a credible outlet repeats your framing, answer engines have another high-quality source to cite5

5

.

New Performance Metrics Replace Traditional Rankings

Companies now monitor AI citation frequency, brand mentions inside AI-generated answers, share of voice across multiple AI platforms, AI referral traffic, and answer inclusion rates4

4

. Several software platforms already provide dashboards that measure AI visibility instead of keyword rankings, with industry reviews highlighting tools from Semrush, Ahrefs, Surfer, and Mentions as leading options for brands that want detailed insight into AI search performance4

4

. The next phase of digital competition will focus less on page rankings and far more on citation in AI recommendations, authority, and retrieval success inside AI-generated answers4

4

.