Scenario: Google Consolidates, AI Redesigns the Customer Journey

Users still primarily search on Google

But the rise of Large Language Models (LLMs) like ChatGPT has introduced a new paradigm for finding information and products. Instead of presenting a traditional list of links to explore, these systems provide direct, concise, and conversational answers, reshaping user expectations.
Despite this revolution, the current landscape still sees Google in a position of absolute dominance. Data are clear: as of September 2025, Google holds about 90% of the global search market (Statcounter), and its ecosystem is so entrenched that 95% of users who start browsing on ChatGPT also use Google as the next step; conversely, only 14% of users coming from Google visit ChatGPT as a continuation of their user journey (Search Engine Land).
Moreover, Google and ChatGPT only partially overlap in their use cases: Google remains king for quick-action searches — navigation, fast answers, purchases, and local queries. ChatGPT, on the other hand, excels at complex and generative search: synthesizing information, making detailed comparisons, and creating content. While their uses are not equivalent, both platforms are increasingly encroaching on each other’s territory: for example, product evaluation is becoming common ground.

Google’s leadership is no longer guaranteed. To defend its position and neutralize the threat, Google’s countermeasure has been to integrate LLM logic directly into the SERP. This has given rise to AI Overviews, AI-generated summaries that answer complex queries, and AI Mode, which transforms search into a continuous dialogue. In this way, Google is evolving its interface to capture and satisfy, within its own ecosystem, the needs that would otherwise push users toward external platforms, creating a hybrid experience that mirrors the conversational journey typical of LLMs.

LLMs as a Parallel Search Destination

Despite Google’s dominance, the gap with conversational platforms is gradually narrowing. While Google once enjoyed a search volume hundreds of times greater, recent data (August 2025, via Similarweb) show a different picture: out of Google’s over 83 billion monthly visits, ChatGPT accounts for nearly 6 billion. Although the absolute gap remains large, it has narrowed, indicating a radical shift in user behavior.
ChatGPT has established itself not as a direct competitor in total volume, but as a parallel destination for complex and informational searches, capturing about 9% of all digital queries (First Page Sage, Q2 2025).
The same study highlights that transactional queries have steadily grown on ChatGPT from Q1 2023 to Q2 2025, reaching 5% in Q2. Over 91% of e-commerce product searches now generate AI-driven results (via Prerender, October 2025), with coverage reaching 94–95% in the fashion and beauty sectors.
The purchase journey is evolving, along with the type of search results and the way users interact with them, with clear implications for e-commerce. Next, we trace the different phases of LLM and generative AI usage.

LLMs: From Discovery and Inspiration Engine to Decision Support and Transactional Interface

1) Discovery and Inspiration Engine

In the early phase of LLM adoption, users ask broad, generic, and open-ended questions to get ideas or find inspiration for a product or service choice.

Example queries:

• “What are the best gift ideas for a wedding anniversary?”
• “Explain the advantages of an air fryer compared to a traditional oven.”
• “I’m looking for a multi-day trekking backpack; what features should I consider?”

LLMs act as search assistants, synthesizing information from the web to provide a general overview, without necessarily linking to specific products.

2) Decision Support

When users have a clearer idea, they use LLMs to compare, validate, and refine their choices. Interactions become more specific and targeted, such as requesting direct comparisons and personalized recommendations.

Example queries:

• “Compare Sony WH-1000XM5 headphones and Bose QuietComfort Ultra based on sound quality and battery life.”
• “Find me a 55-inch 4K TV with at least 3 HDMI ports under €600.”
• “I have sensitive skin; what is the best fragrance-free SPF 50+ sunscreen?”

Here, the LLM becomes a purchase advisor, leveraging structured data (technical sheets, reviews, comparison articles) to provide reasoned, personalized responses that help users make informed decisions.

3) Transactional Interface and Agentic Checkout

Users express a clear purchase intent, view product details, and can complete the transaction.

Example queries:

• “I want to buy the black Sony WH-1000XM5. Add them to my Amazon cart.”
• “Reorder the dog kibble I purchased last month.”
• “Book this hotel in Rome from November 10–12 for two people.”

Since September 2025, ChatGPT can more easily access product information if merchants populate the OpenAI feed, transforming into a checkout interface that communicates directly with e-commerce systems, using stored user information (address, payment method).
With the launch of Instant Checkout and the Agentic Commerce Protocol, OpenAI has radically changed the e-commerce landscape (Wordlift, October 2025). The 700 million weekly ChatGPT users in the U.S. can now purchase products directly through the chat interface without leaving it. AI agents complete purchases on behalf of users, rather than requiring step-by-step manual input (Search Engine Land, October 2025). AI no longer just helps users find external websites to complete actions (e.g., a booking site link) but can execute the action directly, handling search, comparison, and filtering, reducing cognitive load and presenting only the final decision moment (e.g., “Do you confirm the purchase?”).

But it’s not just OpenAI. While ChatGPT was making headlines, Google’s AI mode implemented agentic features capable of searching across multiple platforms, tracking prices over time, and automatically completing purchases via Google Pay when user-defined conditions are met. With access to over 50 billion product listings in its Shopping Graph, updated hourly with 2 billion ads, Google holds infrastructural advantages in data that no competitor can easily replicate.
In summary, users are no longer using AI chat solely for general product discovery. They are integrating it as a true conversational layer in the decision-making process, formulating complex requests and expecting comparative, well-argued responses.
The integration of product feeds and checkout functionality directly into conversational interfaces represents the horizon toward which e-commerce is moving. Being present and influential at this stage of the user dialogue is becoming a crucial competitive factor.

What is GEO (Generative Engine Optimization)?

GEO is the discipline focused on optimizing a brand and its products’ visibility within the responses generated by generative engines, including LLMs, AI Overview, and AI Mode. The goal shifts: it’s no longer just about ranking a link, but about becoming the source from which the AI draws to construct its response.

Strategic objectives of GEO:

• Become an authoritative source: Be the primary reference for the LLM when generating answers for queries relevant to our market.
• Increase brand visibility: Make the brand more visible, increasing the likelihood of being recommended (linked or cited) in LLM-generated responses.
• Control narrative and improve sentiment: Ensure that information about the brand and products is presented accurately and aligned with strategic positioning.
• Drive conversions: Guide users toward the e-commerce site, through citations, links in the generated text, or direct transactional integrations, increasing the channel’s contribution to overall revenue.

SEO and GEO: A Synergistic Approach, Not a Replacement

At Storeis, we see GEO as enhancing SEO, not replacing it. GEO is a strategic layer built on solid SEO foundations. The core principles of SEO—authority, relevance, and quality—become indispensable prerequisites for successful GEO. It’s time to move beyond simplistic narratives and false myths.

False Myth #1: “SEO is dead”
This recurring statement is unfounded for several strategic reasons:

• Market share dominance: Traditional search engines, primarily Google, still hold absolute market dominance. Ignoring standard SEO effectively means giving up the primary source of qualified traffic for almost any online business. Pragmatically, SEO remains critical for measurable results.
• Traditional SERP as primary traffic source: While AI Overview, AI Mode, and LLMs offer direct, low-click access to results, classic snippets remain the battlefield for traffic. AI-driven page visits alone cannot replace the organic clicks from the SERP.
• Correlation with ranking: Analyses of sources used by Google’s AI Overview and AI Mode show a strong correlation with top organic SERP results. Investing effectively in SEO translates into a direct competitive advantage for GEO, as LLMs rely on sources that Google already trusts.

False Myth #2: “SEO and GEO are the same thing”
Although GEO builds on SEO principles, it requires a specific focus on technical and content elements that are critical for machine interpretability.

• AI crawler accessibility and technical optimization: Just like traditional SEO, the technical architecture of the site is crucial. Recent studies (Prerender, October 2025) show that LLMs struggle to correctly index content generated via JavaScript. Technologies such as Server-Side Rendering or Dynamic Rendering are no longer merely SEO best practices—they are essential to ensure that AIs (especially those with less sophisticated crawlers than Google) can accurately read and understand product information.
  
• Structured Data (Schema Markup): Granular and precise markup is now crucial to communicate product information clearly to AI: specifications, price, availability, and relationships with other elements. The ultimate goal is not just to mark data, but to build a Product Knowledge Graph: a rich, interconnected information model that becomes the single source of truth for any AI analyzing the catalog.
• Content Architecture: Content must be formatted for dual readability—human and AI. Heading hierarchies (H1, H2, H3), lists, tables, and clear, unambiguous language are essential for parsing and information extraction by LLMs.
• Focus on Conversational Intent: It’s necessary to map and answer specific, complex questions (long-tail queries) that a user might pose to a voice assistant or chat. Content should serve as the definitive answer to potential conversations (guides, how-to, FAQs…).
• Focus on Conversion: The entire site must be designed to facilitate action completion by automated entities, ensuring that agentic checkout functions correctly.
  

False Myth #3: “Optimization only happens on the owned website”

This is one of the most limiting and dangerous beliefs in the era of generative search. Thinking your website is an island and that optimizing it alone is enough for success ignores how AIs build their knowledge. An LLM doesn’t only read your site; it scans the entire web to form an opinion about your brand and products.
Authority and trust (E-E-A-T) are built across a broad ecosystem of signals. A mature GEO strategy must therefore monitor the entire digital footprint of the brand:

• Forums and communities (e.g., Reddit): Goldmines of authentic conversations. They are often cited as sources in AI Overview. Opinions, doubts, and real user experiences here are extremely valuable signals for AI evaluating product experiences.
• Review sites and UGC content: Social proof is a pillar of trust. Positive reviews on authoritative third-party platforms heavily influence an AI’s perception of a brand.
  
• Social Media (e.g., Instagram): Especially with Instagram indexing all content by default (July 10, 2025, via Seozoom), social platforms become direct sources for search engines. How a brand is represented, what users say in comments, and how products are used all contribute to the overall picture.
  

A strategic approach must be synergistic. Actively monitoring these external touchpoints helps understand brand perception and, where possible, optimize your presence. Most importantly, insights gathered from these conversations (recurring questions, criticisms, praise) should inform content improvements on the owned website, creating a virtuous cycle that strengthens brand authority across all fronts.

In Conclusion: Building Trust in the Era of Instant Answers

Generative Engine Optimization is not a threat—it is the natural evolution of the digital landscape. At Storeis, we see it as an opportunity for ecommerce brands to build deeper authority and presence, understand their users better, and engage them at increasingly relevant points of their buying journey.
While traditional SEO aimed for visibility, GEO’s objective is trust. In a world where AI synthesizes, summarizes, and recommends, the winning brand is the one AI considers the most reliable source. Being that source is not accidental; it results from a precise strategy combining impeccable technical foundations, structured data for accurate product knowledge, and authority built across the entire digital ecosystem.
Our vision at Storeis is clear: SEO and GEO are synergistic disciplines working toward the same goal—making our clients the definitive answer to their users’ questions, regardless of the platform. From yesterday’s traditional SERP to today’s LLMs, our mission is to ensure that the information guiding decisions is accurate, useful, and controlled. SEO taught us the rules; GEO challenges us to apply them on a new, demanding battlefield.