How to Get AI to Recommend Your Brand: Research Reveals Limits of Technical Fixes

    Traditional search engine optimization methods are falling short in the AI era. Extensive research analyzing over 158,000 AI responses shows that brand recommendations are driven by content depth and relevance to user queries rather than mere technical tweaks.

    MediaInsideMediaInside Editorial Team5 min read
    How to Get AI to Recommend Your Brand: Research Reveals Limits of Technical Fixes
    Share:

    Imagine a customer who isn't searching for a specific brand or product SKU, but needs to solve a real-life problem: finding functional underwear for high-intensity sports, the right anniversary gift for a partner, or jewelry durable enough for everyday wear. While they once browsed search engine links, today they increasingly turn to artificial intelligence. AI doesn't just offer specific products—it explains their benefits right away. But how do you make this shortlist? New research suggests that commonly recommended technical tricks simply aren't enough.

    A recent analysis by researchers Mahe Chen and Nick Koudas draws on data from 158,099 AI responses and 468,037 citations. Published as an arXiv preprint on October 10, 2026, the study challenges long-held assumptions about brand visibility in the generative search era. As it turns out, merely adding an FAQ section or deploying structured data won't automatically convince an AI system to pick your online store.

    AI citations depend primarily on a page's relevance to the user's specific question and how the system retrieves it.

    GEO: The New Challenge for E-Commerce

    With the rise of artificial intelligence, marketer attention is shifting toward GEO (Generative Engine Optimization)—optimizing visibility in systems that synthesize comprehensive answers rather than returning simple lists of links. The research revealed that the primary driver of citations is how well titles and accompanying copy match real user queries. Summary boxes or underlying website code alone did not account for differences in citation rates between comparable pages.

    In controlled experiments, the researchers also uncovered an intriguing phenomenon: changing the brand name altered the ranking of recommended products, even when all other product specifications remained identical. This means recommendations are influenced not just by freshly retrieved content, but also by what the model learned about the brand during training. For online retailers, this demonstrates that established brand recognition plays a far larger role in digital recommendations than previously assumed.

    Turning Your Website into an Answer for AI

    Retailers face a fundamental question: does their online store provide the information shoppers actually need to make decisions? Vague claims of "high quality" are unhelpful to both algorithms and humans. Experts recommend taking cues directly from customer support. Recurring questions are the best indicator of missing product details—whether that means omitted dimensions, specific material properties, or clear comparisons between variants of a single product.

    Based on research findings and practical experience, e-commerce stores should focus on the following areas:

    • Concrete answers — Copy must directly address real customer needs (e.g., choosing materials for sports).
    • Brand clarity — The brand must clearly convey what it offers and how it stands out from competitors.
    • Contextual relevance — Product descriptions should explain how an item is used or paired in scenarios photos cannot convey.
    • Monitoring search queries — Continuously test how various AI assistants respond to common questions about your product range.

    Changing a brand name shifted product rankings in AI systems, even when all other features stayed identical.

    The Role of Copy When Photos Aren't Enough

    For categories where emotion and aesthetics lead the way, such as jewelry, written descriptions play an even more vital role. Shoppers browsing online cannot try an item on or gauge its true weight. Nicole Váchová of the specialized store franco bene confirms that accurate copy must supply what visuals miss—scale, wearability, or material specifics. These precise details assist both the customer making a purchase and the AI answering a query.

    “In my view, the research shows that AI visibility cannot be solved with a single technical fix. What matters is whether a store provides a concrete answer to what the customer is trying to solve. For e-commerce businesses, this is an important takeaway: alongside search engine rankings, we also have to consider how our products are understood by systems guiding people's choices,”

    Petr Polák, Trenýrkárna.cz online store

    Ongoing Work Instead of One-Off Tricks

    Chen and Koudas's study does not imply web owners should abandon technical maintenance or structured data. Technical elements still serve a purpose, but they are not an automatic ticket to a recommendation. The path to AI visibility lies in continuous content refinement and data precision. It is also crucial to distinguish between AI merely mentioning a brand versus actively recommending it as the answer to a specific need—a distinction with very different business outcomes.

    Retailers like Trenýrkárna.cz and franco bene are already adapting to these trends by systematically restructuring their sites. They are enhancing content around real questions shoppers ask regarding cuts, fits, and fabrics. This is an ongoing process that doesn't end with a site launch; it requires continuous monitoring of how and in what contexts AI presents their products. After all, there is still no silver bullet.

    AI Analysis

    AI

    Automatically generated commentary • Does not replace expert analysis

    Let AI analyse this article — market impact, comparison and historical context.

    Early Signal Detector

    AI

    Captures signals before the mainstream — inspired by Dataminr

    Reveal early signals and trends the mainstream has not yet captured.

    Context Connector

    AI

    Historical parallels, similar cases and hidden patterns

    Discover what connects to this article — historical parallels, similar cases and cross-industry insights.

    MediaInside
    MediaInside Editorial Team

    Editorial Team

    Under the brand MediaInside Editorial Team, texts are published that are the result of collective editorial work – summaries, short news, overviews, and comments, where the entire editorial team, not one specific author, is responsible for the content.