How AI is changing SEO

For the past two decades, classic SEO has been one of the main sources of traffic on the internet. In 2024–2025, however, this order began to wobble: users increasingly receive answers directly from language models—in Google (AI Overviews / SGE), in ChatGPT, Gemini, Perplexity, or Copilot—instead of clicking on links.

Against this backdrop, a new term has emerged: Generative Engine Optimization (GEO). It is positioned as a response to declining search traffic and a way to gain “visibility in the AI era.” But is GEO truly a new source of traffic, or rather an attempt to salvage what AI is taking away?

What are “generative engines,” and where did GEO come from?

Generative engines are systems that generate an answer to a user query instead of presenting a list of links. These include, among others:

Google Search with AI Overviews / SGE,

ChatGPT, Gemini, Claude, Copilot,

answer-focused search tools such as Perplexity AI.

In 2023, a group of researchers (Gao, Liu, Si, Meng, Xiong, Lin) introduced the term Generative Engine Optimization (GEO) as a new paradigm for content optimization—not for classic SERPs, but for visibility in AI-generated answers.

The key difference:

SEO – you compete for position in Google/Bing search results.

GEO – you compete to have your brand or content cited or mentioned in AI-generated answers.

Today, GEO is discussed both in industry literature and in Polish educational materials and agency offerings (Deloitte, Senuto, Sempire, Verseo, Whites, etc.).

Why did GEO emerge at all? Traffic declines and “zero-click searches”

The trigger is the rapid growth of so-called zero-click searches—situations where users receive an answer directly in the results (or via an AI tool) and do not visit a website.

Industry reports point to, among other things:

estimated organic traffic declines of around 30–40% following the introduction of AI Overviews / SGE, alongside an increase in query volume and CTR drops of even several dozen percent,

a rise in zero-click searches in the news segment and a drop in visits to news portals from over 2.3 billion to below 1.7 billion per month within a year.

In other words, traffic is not disappearing—but an increasing share of it is being “captured” by the AI layer.

This is why GEO is not merely a buzzword, but a response to a real problem: how to recover at least part of the traffic that AI retains?

What exactly is Generative Engine Optimization?

In practical terms:

GEO is the process of optimizing content and online presence so that they are visible, cited, and recommended in AI-generated answers (ChatGPT, Gemini, Perplexity, AI Overviews, etc.).

Typical GEO elements found in analyses and guides include:

Content “to be cited,” not just to rank

Long, in-depth articles with clear definitions, lists, and FAQs.

Explicit answers to user questions that AI can easily quote.

Strong brand and credibility (E-E-A-T) Generative engines are more likely to reference sources perceived as expert, up-to-date, and trustworthy.

AI-friendly structured data and meta-information

schema.org, FAQPage, Product, Organization,

emerging experiments such as llms.txt with instructions for models. ** Visibility analysis in generative engines**

simulating queries in Perplexity, ChatGPT, Gemini,

checking whether and how AI mentions a given brand.

Continuous testing and optimization Startups are even emerging that simulate tens of thousands of prompts and measure brand share in model responses—such as Azoma, which explicitly labels this as GEO and has raised VC funding for it.

Is GEO really a new source of traffic?

The answer is less romantic than sales decks suggest.

1. GEO as “traffic from AI”

New types of traffic sources include:

Clicks on links in AI answers—for example, references in AI Overviews or Perplexity.

Growth in branded searches after AI exposure—users see a brand in an answer and later search for it directly in a browser or store.

Traffic beyond the classic browser—recommendations in Copilot, Gemini on Android, voice assistants, etc., which may lead to app installs or company contact rather than a website visit.

This is no longer just SEO → click → website. Some user journeys take place entirely within the AI ecosystem, with the website becoming just one of several touchpoints.

2. GEO as a “brake on decline”

On the other hand, GEO is largely a defensive strategy:

generative answers have already taken part of organic traffic,

GEO allows you to recover a portion of what was lost—but not necessarily to build something larger than before.

In practice, GEO is therefore:

a new visibility channel,

but the traffic source is hybrid—a mix of SEO, branding, content, and AI integrations.

GEO and the job market – new roles and skills

As GEO becomes a service offered by agencies and the subject of specialized reports, a natural question arises: how will this affect the job market?

1. New specializations at the intersection of SEO, data, and AI

Likely (and partly already visible) directions include:

GEO / AI Search Specialist Combines classic SEO with an understanding of how LLMs and generative search interfaces work. Responsible for testing brand visibility in ChatGPT, Gemini, and Perplexity, content optimization, and competitive benchmarking.

Answer Engine Optimization (AEO) Consultant Focuses on making a company the “first answer” in AI tools—especially in B2B, legal, medical, and similar niches.

Data / Content Engineer for GEO Handles the technical layer: structured data, APIs, headless CMS integrations, optimization for crawling and indexing in the AI context.

For SEO specialists, this means the need to upskill in AI/ML; for engineers, greater involvement in designing content systems and visibility analytics for generative engines.

2. What does this mean for developers?

For developers—especially web, backend, and DevOps—GEO brings concrete tasks:

Implementation of structured data and APIs

schema.org, JSON-LD,

API endpoints with “clean” data for integration with external systems and assistants.

Performance and availability Even the best GEO content won’t help if a site is slow or unstable—models and crawlers may deem it a less reliable source.

Logs and analytics “beyond Google Analytics” Teams need to learn how to interpret traffic from new sources (e.g., referral tabs showing non-obvious hosts, campaign parameters from AI assistants, growth in direct traffic after AI exposure).

Building internal “mini answer engines” Companies investing seriously in GEO often build internal RAG/LLM solutions in parallel to better understand how models select sources—another field for Python/ML/DevOps.

In short, GEO is yet another reason for developers to go deeper into AI and search systems, not just “raw code.”

3. Agencies and software houses – new services, new teams

Already visible in Poland today:

GEO service offerings alongside SEO/SEM,

educational content explaining the differences between SEO and GEO, aimed at e-commerce and B2B clients.

This translates into demand for:

SEO specialists with AI skills,

content marketers who can write “for citation” by models,

technical specialists implementing structured data and integrations.

For someone just entering the IT job market, GEO can be a way to stand out—combining SEO knowledge with an understanding of how LLMs work is still rare.

GEO in practice – what does “optimization for generative engines” look like?

In simplified form, a typical GEO strategy (based on English-language frameworks and guides) includes several steps:

1. AI presence audit

checking what answers ChatGPT / Gemini / Perplexity provide for key questions in your industry,

noting whether your brand or site is mentioned.

2. Topic and intent mapping

Not just keywords, but entire user question scenarios—from “what is GEO” to “which agencies in Poland offer GEO.”

3. Designing “AI-first” content

expert-level, context-rich content,

clear headings, definitions, lists, FAQs,

language understandable for humans but well structured for models.

4. Structured data and credibility signals

schema.org markup, clear author attribution, update dates,

links from reputable sites and presence in trusted directories and industry portals.

5. Monitoring and iteration

re-testing AI tools after changes,

analyzing gains and losses in mentions,

experimenting with different content formats (case studies, reports, white papers).

This directly generates tasks for IT/data teams—from tools that automatically check model responses to integration of logs and reporting systems.

Will GEO “replace” SEO?

Some marketing materials eagerly proclaim “SEO is dead, GEO is coming.” Reality is less dramatic.

A sober comparison:

SEO will remain essential—generative engines rely on classic quality signals: content, links, domain authority, technical site health. Poor SEO = a weak foundation for GEO.

GEO is rather a layer on top of SEO—optimizing how AI cites your content, not whether it sees you at all.

Traffic from classic Google will likely shrink, but not disappear—many query types (transactional, navigational, local) will still generate link clicks, even in an AI-driven world.

For the job market, this means not the disappearance of SEO specialists, but an evolution of roles. Just as the “webmaster” role once split into frontend, backend, DevOps, and product roles, today the “SEO specialist” is beginning to fragment into:

technical SEO/GEO,

GEO content specialist,

AI search analyst.

Summary: GEO as a new front in the battle for attention, not a miracle traffic source

To the question “Generative Engine Optimization – a new source of traffic?”, the answer would be:

Yes, but… GEO opens up a new visibility channel—in generative answers, AI recommendations, and assistants—and makes it possible to generate traffic that did not exist before (e.g., clicks from Perplexity, exposure in Copilot).

…and at the same time, not entirely. To a large extent, GEO is an attempt to recover lost organic traffic that AI layers have absorbed. Without solid SEO and high-quality content, GEO will remain little more than a buzzword on slides.

For the IT job market, this means:

expanding classic SEO with AI and analytics skills,

new roles at the intersection of development, data, and marketing,

opportunities for specialists who are first to learn how to think about visibility **from the perspective of models, not just search engines. ** In other words, generative engines are not ending the traffic game—they are simply moving it to a different board. Those who learn how to play on it won’t complain about “the end of SEO,” but will calmly add a new line to their business card: GEO / AI Search Specialist.