AI-assisted merchandising The use of artificial intelligence to decide which products to show, in what order, and to whom, without relying on manual rules for every scenario. Definition Definition AI-assisted merchandising uses machine learning to decide which products get shown, in what order, and to which shopper, instead of relying on a person writing a rule for every possible scenario. Traditional merchandising runs on rules set by hand: "when someone searches trainers, show brand X first," or "in the coats category, push higher-margin products to the top." Rules like these give a team direct control, but they don't scale. A catalog of 50,000 products, combined with hundreds of possible search terms, seasons, stock levels, and shopper profiles, creates more scenarios than any team can realistically cover with manual rules. AI-assisted merchandising works from real-time signals instead: what each shopper searches for, which products they view, what they add to cart, what they skip, and what shoppers with similar behavior end up buying. Based on that, the system adjusts the order of search results, category pages, and recommendation carousels on its own, without someone editing a rule every time something changes. That doesn't remove human control from the picture. Ecommerce teams still set the strategy, like prioritizing clearance stock, pushing a new collection, or protecting margin. The system just handles execution at a scale no team could do manually, product by product, shopper by shopper. What it's used for What it's used for The most direct benefit is relevance at scale. A shopper searching for "moisturizer" in a cosmetics store shouldn't see the same results as someone who has made three visits comparing anti-aging products. Without AI, both see the same list. With AI, each sees a different order, aligned with their purchase intent. In practice, this means fewer zero-result searches, lower bounce rates from category pages, and higher average order value, because complementary product recommendations are based on real co-purchase patterns rather than generic associations. The less visible but equally concrete benefit is operational time savings. Merchandising teams stop maintaining dozens of manual rules and can focus on strategic decisions the AI can't make for them: brand positioning, campaign narrative, or selecting new suppliers. Doofinder Doofinder combines manual control (searchandising) with AI-driven personalization. The system collects user behavior data — pages viewed, clicks, products added to cart, and purchases — and uses it to adjust search result rankings and product recommendations in real time for each visitor. Specific features include: 1:1 personalized results: Search results are reordered based on each user's behavioral profile. Two people searching for the same term see different products based on their previous interactions. AI-powered recommendation carousels: Product carousels on the homepage, product pages, and category pages are driven by behavioral analysis and image matching to suggest visually similar or complementary products. Automatic visual tagging: The AI analyzes catalog images to generate attribute tags (color, style, material), improving search accuracy and enabling merchandising to draw on metadata that would otherwise need to be entered manually. Searchandising: eCommerce teams can create boosting rules, pinning rules, and promotional banners within search results, and the AI complements these by optimizing the rest of the ranking. This hybrid model — manual rules plus AI — allows the team's strategic decisions to coexist with the kind of automatic optimization that only machine learning can execute at scale. Example Case study A sportswear retailer with 8,000 products launches its summer collection. The merchandising team creates a rule to ensure new products appear in prominent positions when someone searches for "running" or "trail running." That's manual merchandising. The AI complements that rule by detecting that users who have bought trail running shoes in the last 30 days tend to be interested in technical socks and hydration vests. For those profiles, it reorders results to intersperse accessories among the shoes. For a new visitor with no purchase history, it maintains the standard order set by the team.