Northcroft Outdoor, a direct-to-consumer retailer specializing in high-end hiking and camping equipment, built its digital presence on a combination of Google Shopping, long-tail SEO content, and a loyal email list. By mid-2025, the marketing director noticed that customer acquisition costs were rising while conversion rates on paid traffic were declining. A segment of new customers arriving through organic channels was converting at twice the rate of paid traffic -- and when the team surveyed these customers, a consistent pattern emerged: many had discovered Northcroft after asking an AI assistant what gear to buy.
The problem was that this discovery was accidental. Northcroft's products appeared in AI responses only when a customer asked a very specific question that happened to match existing content. For broad category queries -- "best trail running backpacks," "most durable trekking poles under $200" -- the brand was absent. Competitors with inferior products but better AI-optimized content infrastructure were consistently recommended instead.
AISOS was engaged to systematically rebuild Northcroft's AI visibility across product categories. The engagement covered technical implementation, content restructuring, and product data optimization for machine consumption. The results within one quarter reshaped how the team thinks about digital acquisition. For context on the broader approach, see our guide to Answer Engine Optimization.
The Challenge
Northcroft's catalog contained 840 SKUs across 12 product categories. The website had been built for SEO performance: detailed category pages, long-form buying guides, and a review integration that surfaced verified customer ratings. Despite this infrastructure, an AI visibility audit conducted at the start of the engagement revealed that Northcroft appeared in fewer than 8% of AI-generated responses to category-level shopping queries.
The core issue was that AI assistants use different criteria than search engines to select products for recommendation. Search engines rank pages. AI assistants synthesize information and recommend specific items based on their training data and the sources they query in real time. Northcroft's product pages were structured for search crawlers but not for language model comprehension. Product descriptions focused on features and specifications rather than use-case matching. Comparison content was absent. Schema markup was limited to basic Product and Organization types without the richer attributes LLMs use to match products to queries.
The competitive disadvantage was compounding. Two of Northcroft's main competitors had significantly better AI presence despite having smaller review counts and lower domain authority scores. They appeared consistently in Perplexity responses and in ChatGPT's "shopping assistant" mode. Understanding AI visibility as a distinct discipline from SEO was the conceptual shift the team needed before implementation could begin.
The AISOS Strategy
Given the scale of Northcroft's catalog, the implementation strategy prioritized the top 80 SKUs by revenue contribution -- roughly 10% of products generating 62% of revenue. This allowed the team to concentrate AI visibility investment where the commercial impact would be most measurable. The remaining catalog would benefit from the technical infrastructure improvements deployed across the site, but the content work was concentrated on the priority tier.
The content restructuring for each priority SKU followed a standardized template developed by AISOS: a clear use-case definition in the first 50 words, explicit comparison language referencing the specific scenarios where the product outperforms alternatives, and factual specification claims formatted for easy extraction by AI systems. This approach aligns with AEO principles for product content. Product schema was expanded to include offer conditions, availability regions, and review aggregates in JSON-LD format on every product page.
In parallel, AISOS deployed 14 use-case-specific buying guides targeting the exact question formats that AI assistants receive for outdoor gear. These guides were not SEO articles repurposed with minor edits -- they were purpose-built for machine reading, with explicit recommendation logic, comparison tables in structured format, and clear attribution of why specific products suit specific users. The AI SEO checklist guided the technical validation of every page before publication. Links to the e-commerce AI visibility strategy provided additional framework for the category approach.
The Results
At the end of the first quarter post-implementation, Northcroft's AI mention rate on priority category queries had increased from 8% to 54%. The most dramatic improvement was in Perplexity's shopping recommendation responses, where Northcroft went from appearing in 1 of 15 product queries to 10 of 15. ChatGPT's browsing-enabled responses showed an increase from 2 of 15 to 9 of 15. Claude showed 7 of 15, up from 1.
Revenue attributable to AI-sourced discovery increased by 41% quarter-over-quarter. This was measured through a combination of UTM tracking on AI platform referral traffic (which is trackable when users click through from Perplexity), post-purchase survey data, and a "how did you first hear about us" question added to the checkout flow. Average order value from AI-sourced customers was 18% higher than the site average, consistent with the hypothesis that customers who arrive via AI recommendation are further along in the purchase decision and more likely to buy premium-tier products.
The buying guides published as part of the engagement accumulated 23,000 organic search impressions in the first month -- a secondary benefit not originally targeted. The structural quality required for AI readability also produced content that performed well in traditional search, demonstrating that the two approaches are compatible when implemented correctly.
Key Success Factors
Prioritization was the single most important decision in this engagement. Attempting to restructure all 840 SKUs simultaneously would have diluted resources and delayed meaningful results. By focusing on the top 80 revenue-generating products first, the team achieved measurable AI visibility improvements within 60 days -- fast enough to maintain organizational momentum and justify continued investment in the broader catalog.
The buying guide strategy was effective because it matched the exact format of questions AI assistants receive from users. Generic content about outdoor gear generates generic AI responses that mention many brands. Highly specific content -- "which backpack is best for multi-day alpine routes with a starting budget of $250" -- gives AI systems the precise, citation-worthy answer they need. Specificity drives citation, and citation drives discovery. This is a core principle of AEO applied to e-commerce.
Integration between the content and technical teams was essential. AISOS structured the engagement so that schema deployment and content updates were synchronized at the page level -- meaning every restructured product page was simultaneously updated with expanded schema and revised copy. Decoupled implementations where content and technical changes go live at different times lose the compounding benefit of simultaneous signal improvement. Teams considering a similar engagement should ensure their CMS allows synchronized publishing across both content and metadata layers.
Lessons Learned
The most counterintuitive finding from the Northcroft engagement was that product specification data -- the dry, unglamorous tables of dimensions, weights, and materials that most content teams deprioritize -- was among the most powerful AI visibility signals. Language models love specific, verifiable facts. A product page that precisely states "total pack weight 1.24 kg, capacity 38 liters, frame-to-hipbelt distance adjustable 42-54 cm" gives an AI system the exact data it needs to match the product to a user's query. Feature-marketing language ("revolutionary comfort system, industry-leading durability") gives the AI model nothing it can work with.
The second lesson is about the review integration. Northcroft had 14,000 verified customer reviews on their website. None of them were structured in a way that AI systems could sample. Reviews buried in a JavaScript widget with no static HTML rendering are invisible to LLM crawlers. Restructuring the review display to include static HTML with schema markup immediately made this asset useful for AI visibility purposes. If your brand has significant review volume, rendering it in machine-readable format is one of the highest-leverage actions available.
Finally, the engagement reinforced that AI visibility is a channel that rewards consistency over time. The models that showed the fastest improvement (Perplexity, with its real-time web access) responded within weeks. The models with slower corpus update cycles (Gemini, Copilot) continued improving month over month. Brands that start early benefit from this gradual accumulation effect. Reach out to AISOS to understand the current state of your AI visibility and the fastest path to improvement.