The Agentic Shift: Why Myer’s Parabolic LLM Traffic Growth is a Wake-Up Call for Global Retail

The era of "browsing" is being disrupted by "fetching" as AI agents take the wheel of the consumer journey. For CMOs and e-commerce leaders, the recent data from Myer proves that Generative Engine Optimization (GEO) is no longer a theoretical future—it is a current, high-conversion revenue driver that requires a fundamental rethink of product data architecture.

For decades, the e-commerce playbook was built around the search box. Whether it was Google, Amazon, or an on-site search bar, the goal was to capture a user’s attention during their manual discovery phase. However, the latest reports from Myer, one of Australia’s largest department store chains, suggest we have hit a tipping point. Myer is seeing what they describe as "parabolic growth" in traffic originating not from humans clicking links in Google, but from Large Language Models (LLMs) and autonomous AI agents.

What makes this shift particularly jarring for traditional marketers is the quality of the traffic. According to the data, these AI-driven interactions are converting at rates significantly higher than traditional paid search or social media. This is because an AI agent—like a custom GPT, a Perplexity user, or a shopping-specific LLM—only directs a user (or a purchase) to a site when there is a high-confidence match between the consumer’s complex intent and the product’s specific attributes.

Simultaneously, the infrastructure for this "Agentic Commerce" is being formalized. Mirakl, the platform powering many of the world's largest retail marketplaces, has launched "Agentic Activation" tools. These tools are designed to ensure that product catalogs are not just visible to human eyes, but are structured for "machine-readability." This allows autonomous agents to find, evaluate, and even execute purchases on behalf of the consumer. We are moving from a world where we market to people, to a world where we must market to the algorithms that represent them.

The strategic shift here cannot be overstated: visibility is moving from the search results page (SERP) to the "latent space" of an LLM’s training data and real-time retrieval systems. When a consumer asks an AI, "Find me a sustainable winter coat under $300 that fits a minimalist aesthetic and can ship to Melbourne by Friday," the AI doesn't give them a list of ten blue links. It provides a recommendation based on the data it can most easily ingest and trust.

For brands, this means the "moat" is no longer just brand equity or high ad spend; it is the precision and accessibility of your product data. If an AI agent cannot verify your stock levels, your material composition, or your shipping speed through a structured data format, your brand effectively does not exist in the agentic economy. This is the essence of Generative Engine Optimization (GEO).

Furthermore, the higher conversion rates reported by Myer highlight a reduction in "friction-based browsing." In the traditional model, a user might click five ads before finding what they want. In the agentic model, the agent does the "clicking" in the background. By the time the consumer sees your brand, the intent is almost fully realized. This requires marketing leaders to shift their KPIs. We are no longer just looking for "clicks"; we are looking for "agentic mentions" and "recommendation share." If your inventory isn't machine-readable, you are ceding the most valuable, high-intent traffic to competitors who have optimized for the LLM era.

To stay competitive in an environment where AI agents make the decisions, brands must move beyond traditional SEO and adopt a GEO-first mindset.

• Action 1: Prioritize Extreme Schema Markup and Structured Data: You must go beyond basic "Product" schema. Implement deeply nested JSON-LD that includes real-time availability, detailed material provenance, specific use-case scenarios, and compatibility data. AI agents rely on these structured "clues" to categorize your products within their reasoning frameworks.

• Action 2: Develop "Agent-Friendly" Product Descriptions: Traditional product copy is written to evoke emotion in humans. While still important, you now need a layer of data that satisfies LLM logic. Use clear, descriptive, and non-ambiguous language that defines the "utility" of the product. Describe the "why" and "how" of the product in a way that helps an LLM understand its context for complex natural language prompts.

• Action 3: Open the "API Gates" for Real-Time Retrieval: LLMs are increasingly using "tools" or "plugins" to browse the live web (e.g., OpenAI’s GPT-4o or Perplexity’s real-time search). Ensure your site’s internal search API or product feeds are accessible and optimized for fast, accurate retrieval by these crawlers. A "headless" commerce approach that prioritizes data delivery over visual rendering is a significant advantage here.

• Action 4: Monitor "Referral LLM" Traffic in Analytics: Start segmenting your traffic by LLM referrers. While some AI agents mask their identity, many are identifiable. Track the conversion rate of users coming from ChatGPT, Claude, or Gemini. Understanding which agents are currently recommending you will help you understand where your data is most "readable" and where it is failing.

• The Search Paradigm has Shifted: Traffic is moving from the "keyword box" to the "intent prompt." If you aren't optimized for the prompt, you are invisible.
• Data is the New Creative: In agentic commerce, the quality, structure, and machine-readability of your product data are more important than your hero banners or social media aesthetic.
• Conversion is King: AI-driven traffic converts better because the "filtering" happens before the user ever lands on your site. This makes GEO one of the highest-ROI activities for modern e-commerce teams.

To see how your brand currently ranks across ChatGPT, Claude, Gemini, and Perplexity, get a free AI visibility scan at ahod.co/dashboard.