How AI changing fashion product discovery

How is AI changing product discovery in fashion and luxury and what fashion companies should work on

Interview with Giulio Salvucci Digital Fashion expert and founder of Search Bridge.

The interview features a discussion between Enrico Fantaguzzi and Giulio Salvucci, an e-commerce expert and founder of the startup Search Bridge. They discuss how Large Language Models (LLMs) like ChatGPT are fundamentally changing how consumers search for products and what fashion brands must do to adapt to this new landscape.

Key Topics

The Shift from SEO to GEO (Generative Engine Optimization): While traditional SEO focuses on optimizing a single website for search algorithms, GEO requires building brand authority across the entire web. LLMs do not just look at a brand’s website; they synthesize answers by pulling “multi-signal” data from diverse touchpoints like journalistic articles, forums, communities, and review sites like Trustpilot.
The Collapsing Marketing Funnel: The traditional user journey of researching and filtering products across multiple websites is condensing into a single step within AI chatbots. Users can now ask an LLM for specific items (e.g., “a spring sweater in my size and price range”), and the AI will filter catalogs and provide highly targeted recommendations directly in the chat, turning the brand’s website primarily into a final checkout destination.
The Attribution Measurement Gap: Currently, traditional tracking systems rely on “last click” models, making it very difficult to measure how much traffic truly originates from AI searches. Traffic influenced by LLMs often shows up as “direct” or “organic” in analytics, creating a “black box” scenario where multiple customer interactions happen invisibly within the AI.
The Social Media “Blind Spot”: Major social platforms like Meta (Instagram) and TikTok actively block LLMs from scraping their data to protect their own traffic. As a result, massive marketing budgets spent on these platforms do not contribute to a brand’s visibility or authority on generative AI engines.
Content Optimization for Machines: To be cited by LLMs, brands need to adapt their copywriting. Content must be highly factual, clear, and simple—ideally matching the reading level of a 10-11 year old, which can be measured by metrics like the Flesch score. Formats like FAQs and comparative tables are highly favored by LLMs.

Key Takeaways for Brands

Maintain Absolute Consistency: It is crucial that brand information (e.g., number of stores, technical details) is identical across every single online touchpoint; inconsistencies cause LLMs to view the information as unreliable.
Prepare for Direct AI Feeds: In the near future, brands will be able to submit their product catalogs directly to LLMs (similar to Google Shopping feeds), bypassing the need for the AI to crawl their websites.
Manage Online Reputation Proactively: Because LLMs read external sites to form answers, brands must actively engage in digital PR, moderate complaints on sites like Trustpilot, and participate in online community discussions.
Do Not Discount Google: While OpenAI and ChatGPT are popular, Google remains incredibly powerful because it can combine its new AI models (Gemini) with its massive existing infrastructure, including Google Maps and Google Shopping.
Invest in Continuous Learning: The AI landscape is evolving exponentially. Professionals must dedicate significant time to continuous education, as disruptive new technologies can emerge and change the rules in just a matter of months.

How can brands improve their authority to rank in GEO?

To improve authority and rank in Generative Engine Optimization (GEO), brands must shift their focus from purely optimizing their own websites to building a comprehensive, multi-channel digital presence. Large Language Models (LLMs) synthesize answers by drawing correlations from various sources across the web, meaning a brand’s authority is determined by its overall digital footprint rather than just its website.

Here are the key strategies brands should implement to improve their GEO authority:

Generate “Multi-Signal” Traces Across the Web: Brands must establish semantic traces across diverse touchpoints, including journalistic articles, reference sites, forums, and online communities.
Focus on Accessible Platforms: Because major social networks like Meta (Instagram/Facebook) and TikTok actively block LLMs from scraping their data, brands should invest in platforms that AI can actually read, with networks like LinkedIn and YouTube becoming increasingly important.
Prioritize Content Clarity: Content must be highly factual, direct, and objective. To ensure LLMs can easily read and cite the information, texts should be written simply—ideally matching the reading level of a 10-11 year old, which can be evaluated using a Flesch score.
Maintain Absolute Consistency: For LLMs to consider information reliable, it must be identical across every single touchpoint. For example, if a brand has 100 stores, that exact number must be reflected uniformly on its website, press releases, and external articles; any inconsistency will cause the AI to deprioritize the information.
Adopt AI-Friendly Formats: LLMs strongly favor specific content structures. Brands should incorporate FAQs (Q&A mode) and comparative tables, as these formats are easily digested and perfectly read by generative systems.
Proactively Manage Online Reputation: Because LLMs analyze external discussions to gauge sentiment, brands must actively participate in digital PR and online communities. Engaging with customers and resolving complaints on review platforms like Trustpilot significantly improves the brand’s sentiment perception and visibility to the AI.
Prepare for Direct Catalog Feeds: Looking ahead, brands should prepare to submit their product catalogs directly to LLMs—similar to how Google Shopping feeds operate—allowing the AI to access product data locally without having to crawl the brand’s website

How does AI affect the traditional e-commerce marketing funnel?

The traditional e-commerce marketing funnel is essentially collapsing into a single step within AI search engines. Rather than following the conventional journey where a consumer visits multiple websites and manually uses search filters, the AI is now condensing the research and filtering processes into one concentrated channel.

Consumers can simply ask an AI chatbot for highly specific items—such as a spring shirt with a particular pattern, style, or weight. The AI then acts as the filter, navigating various catalogs to present a narrow, curated selection of products from different brands that perfectly match the request.

This transformation affects the funnel in a few key ways:

Accelerated Journey to Checkout: Traditional funnel stages like product discovery, research, and side-by-side comparison are now handled entirely within the AI’s chat interface. This allows the user to get almost to the final checkout stage without ever leaving the AI platform.
Loss of Visibility (The “Black Box”): Because all of this research and consideration happens internally within the Large Language Model (LLM), brands lose the ability to track the early stages of the customer journey. A customer might have several interactions with the AI to narrow down their choice, but the brand only sees the final result when the customer suddenly arrives on their website to make a purchase, turning the top and middle of the funnel into an invisible “black box”

What is a Flesch score and how is it calculated?

The Flesch score is a metric that evaluates the complexity and readability of a text, determining how easily it can be read by both humans and Large Language Models (LLMs). For content to be highly optimized for AI systems, it should achieve a score that reflects a very simple and direct writing style, ideally matching the reading comprehension level of a 10 to 11-year-old child.

The score is calculated by analyzing two main structural elements of the text:

Word length: It measures the number of letters within each single word.
Sentence and paragraph length: It counts the number of words used to form paragraphs.

By keeping words short and paragraphs concise, brands can improve their Flesch score and ensure their content is more easily processed and cited by generative AI engines

How can I learn more about AI in Fashion?

You can learn more about AI in fashion through the Digital Fashion Academy, which offers several live online courses and certifications tailored to different career levels, goals, and time commitments.

Here are the primary learning paths they offer:

1. AI for Fashion Executive Master This is a comprehensive, 14-week part-time online program requiring about 3.5 hours of study per week. It is designed for managers, marketing professionals, and consultants who want to master end-to-end AI applications across the fashion value chain.

What you will learn: How to implement AI in design and trend forecasting, supply chain and operations, merchandising, ecommerce, marketing, and customer service.
Format: It features hands-on labs, real-world case studies, digital portfolio building, and one-to-one mentoring.
Outcome: Upon completion, you earn an internationally recognized AI for Fashion Professional Certificate. The course fee is €2,150.

2. AI for Fashion Executives If you are a senior leader, CXO, or entrepreneur looking for a fast-track option, this is a 6-hour intensive program delivered over two 3-hour live sessions.

What you will learn: It focuses heavily on executive strategy, teaching you how to identify high-ROI use cases, define an AI roadmap, and secure quick wins across design, marketing, and operations.
Format: Interactive live sessions focused on frameworks and actionable strategy. The program costs €350.

3. Short Courses for Specific Skills If you are looking for highly targeted, practical knowledge, there are shorter courses available for €75 each:

Essential AI Training for Fashion Professionals: Teaches you the principles of EU Artificial Intelligence compliance, how to safely use tools like ChatGPT at work, and how to craft effective, reliable prompts specifically tailored for fashion workflows.
Personal Branding in the Age of AI: Focuses on using AI tools to build content libraries and boost your personal brand with a tone of voice that sounds like you.

Across these programs, you will learn directly from senior industry experts and digital strategists who have worked at top global luxury brands, including Gucci, Valentino, LVMH, Dolce & Gabbana, and Burberry. Both the Master and the Executive programs emphasize moving beyond the hype of AI to focus on practical implementation and real business results

Fashion AI Marketing the new GEO paradigm

What challenges fashion brands need to face?

Here is an analysis of the challenges facing fashion companies, structured using the categories of a Fishbone (Ishikawa) diagram to identify the root causes of the core problem.

Core Problem (The “Head” of the Fish): Navigating AI Disruption and Digital Transformation The fashion industry is facing a massive paradigm shift in how consumers discover products, how operations are managed, and how brands must present themselves online due to the rapid integration of Artificial Intelligence.

1. Measurement & Technology Challenges

Obsolete Attribution Models: Traditional analytics rely on “last-click” models, which are incapable of accurately tracking traffic and influence originating from Large Language Models (LLMs).
The “Black Box” Customer Journey: The traditional e-commerce marketing funnel is collapsing into a single step within AI chatbots. Because users research, filter, and compare products entirely within the LLM, brands lose visibility into the customer journey, often only seeing the final interaction when the user arrives to purchase.
Data Flow and Management: Companies struggle with turning raw data into competitive advantages, requiring new workflows for smart data collection, analytics, and business intelligence.

2. Marketing & Visibility Challenges

The Shift from SEO to GEO: Traditional Search Engine Optimization (optimizing a single website) is no longer enough. Brands must now focus on Generative Engine Optimization (GEO), which requires building semantic traces and authority across multiple external touchpoints, such as journalistic articles, forums, and reference sites.
The Social Media Blind Spot: Fashion companies heavily invest their marketing budgets in platforms like Meta (Instagram) and TikTok. However, these platforms actively block LLMs from scraping their data, meaning this massive investment does not contribute to a brand’s visibility on generative AI engines.
Proactive Reputation Management: Because LLMs source answers from across the web, brands can no longer rely just on their owned channels. They must actively monitor and manage their reputation on third-party community sites and review platforms like Trustpilot.

3. Content & Process Challenges

The Copywriting Conflict: There is a clash between the emotional, highly creative copywriting traditionally favored by luxury brands and the direct, simple, and highly factual writing style required by LLMs. To be effectively cited by AI, content needs to match the reading level of a 10-11 year old (measured by the Flesch score) and utilize structural formats like FAQs and comparative tables.
Absolute Information Consistency: AI engines penalize inconsistencies. If a brand states it has 100 stores on a press release but 98 on its website, the LLM will view the data as unreliable. Ensuring exact consistency across every digital touchpoint is a major operational challenge.

4. Operations & Supply Chain Challenges

Efficiency and Optimization: Brands are challenged to optimize their inventory management, demand forecasting, pricing, and returns management in an increasingly complex retail environment.
Sustainability: Companies face increasing pressure to adopt ethical AI, reduce waste, improve supply chain transparency, and implement circular business models.

5. People & Strategy Challenges

The Exponential Pace of Change: The technology landscape is evolving so fast that disruptive new tools emerge every few months. This creates a severe risk of falling behind, making continuous professional upskilling and training an absolute necessity for fashion professionals.
Governance and Risk Management: Leaders must figure out how to navigate complex legal, ethical, and governance frameworks, including the EU AI Act, data privacy, IP protection, and mitigating AI biases and “hallucinations”.
Moving Beyond the Hype: Executives face the challenge of distinguishing real ROI-generating AI opportunities from industry hype, requiring them to define clear roadmaps, select the right vendors, and secure quick wins across design, merchandising, and operations without disrupting existing workflows