Ecommerce LLM SEO Case Study: From 19 to 250 AI Mentions in 60 Days

Traditional SEO focuses on ranking in Google’s blue-link results — earning clicks by appearing on page one. AI SEO focuses on earning citations, mentions, and recommendations from large language models like ChatGPT, Google’s AI Overview, AI Mode, and Gemini. When a buyer asks ChatGPT “what’s the best VOIP phone for a small office?”, AI SEO determines whether your brand gets named in that answer. The disciplines overlap — both require high-quality, well-structured content — but AI SEO places extra weight on entity clarity, topic authority, and creating content that LLMs can confidently cite without hallucinating. MY IT HUB’s campaign showed that the two reinforce each other: optimising for AI citation simultaneously improved Google Search Console rankings.

AI visibility improvements can appear faster than traditional SEO because LLMs update their knowledge through retrieval-augmented generation (RAG) and index new content frequently via their crawlers. In MY IT HUB’s case, AI mentions grew from 19 to 161 within the first six weeks — a 747% increase before the three-month mark. However, compounding is the key mechanic: the more pages that get cited, the more the brand appears in AI training signals over time, creating an accelerating flywheel. Brands should expect meaningful early signals within 4–8 weeks, with material business impact building from month 3 onward.

ChatGPT uses a combination of its base training data and Bing-powered web browsing to answer queries, which gives it a broad citation surface. MY IT HUB’s 300 ChatGPT cited pages (52.6% of total) reflect the fact that its category pages and product content are well-structured enough for ChatGPT to reference confidently. Gemini’s zero citations are likely due to Google’s more conservative citation thresholds — Gemini tends to cite fewer, higher-authority sources and requires stronger entity signals and structured data to recognise a brand as citation-worthy. AI Mode and AI Overview sit in between, drawing on Google’s index with different ranking algorithms than organic search. Closing the Gemini gap requires specific structured data improvements, featured snippet targeting, and building off-site brand signals that Google’s quality raters recognise.

Domain authority matters less for AI citations than it does for competitive organic rankings. LLMs evaluate content at the page level — they look for clear, accurate, well-structured information on a specific topic. A DA 20 site with excellent VOIP phone buying guides can outperform a DA 60 generalist tech site in ChatGPT citations for VOIP queries, because topical relevance and entity clarity outweigh raw link equity in LLM retrieval. MY IT HUB’s campaign confirms this: despite a stable DA of 20, citation counts grew dramatically. That said, domain authority still correlates with Gemini performance, so parallel link-building efforts do contribute to long-term AI visibility growth — particularly for the most contested queries.

LLMs preferentially cite content that is definitional, comparative, or instructional. In the B2B IT hardware context, the highest-performing formats are product category explainers (what is IP DECT, how does it differ from standard VOIP), buying guides with clear criteria, feature comparison tables between product families, and FAQ-style content that mirrors how buyers phrase questions to AI assistants. Thin product listing pages with minimal descriptive content rarely earn citations. The structural upgrade from a basic category page to an entity-rich, question-answering category hub is what drove MY IT HUB’s citation surge — particularly on ChatGPT and AI Overview, which reward informational depth.

This is a common pattern in AI visibility data and does not signal regression. LLM citation indices are dynamic — platforms rotate which pages they surface based on query variation, freshness signals, and periodic re-indexing. A slight dip in cited pages (e.g. from 507 to 491) alongside a rise in mentions (from 173 to 180) indicates that a smaller set of pages is being cited more frequently and more prominently — which is actually a healthier signal than many pages earning single mentions. The goal is not to maximise raw page count but to build a core cluster of highly-cited authority pages that consistently appear across a wide range of relevant queries.

AI SEO ROI is measured across two layers. The direct layer tracks referral sessions from AI platforms (visible in GA4 under source/medium when ChatGPT or Perplexity send browser traffic), branded search volume uplift, and direct traffic growth — all of which correlate with increased AI mentions. The indirect layer tracks the organic SEO uplift that AI-optimised content generates: impressions, clicks, keyword rankings, and new user acquisition. MY IT HUB’s 34.7% sessions growth and 42.9% new user growth are attributable to the combined effect of both layers. Additionally, AI visibility scoring tools (the ones used in this campaign) provide a quantitative citation share metric, allowing month-over-month ROI tracking even when referral attribution is incomplete.

MY IT HUB is precisely the proof point that AI SEO works for transactional, product-focused sites. The misconception is that LLMs only cite blog articles or editorial content. In reality, LLMs frequently answer product research questions — “what’s the best conferencing system for a 20-person office?”, “are Jabra headsets worth it for call centres?” — and they cite category pages, product pages, and buying guides from e-commerce sites when those pages are structured clearly enough. The key is bridging the gap between pure product listing (not citable) and informational product content (highly citable). For B2B e-commerce specifically, where buyers conduct significant pre-purchase research via AI, this is one of the highest-leverage investments a brand can make in 2026.