AI, product visibility and retail conversions

Retailers are competing in a market where demand is harder to predict, and margin is harder to protect.  

 

BDO figures reported by Reuters showed discretionary retail sales across fashion, homewares and lifestyle falling in April, the weakest performance in a decade outside the pandemic period. 

 

For a sector already carrying higher labour, fulfilment and operating costs, product discovery has become more closely tied to revenue. This pressure arrives as customers are starting to search differently.  

 

More shoppers are beginning with a requirement before they have decided where to buy. They may be looking for a waterproof coat for commuting under £150, an outfit that can arrive before the weekend, or furniture that works in a small flat. Budget, timing and practical constraints are already part of the request before the retailer has entered the journey. 

 

Recent moves from John Lewis and Frasers Group show how quickly retailers are adapting to this behaviour. John Lewis is preparing for products to appear in AI environments, while Frasers has launched a conversational shopping assistant. 

 

What sits underneath those developments is a broader shift in how products are discovered. More of the first comparison can now happen before the customer reaches a retailer’s own site. In practice, that means product information has to work harder and earlier in the buying journey. 

 

 

Product data starts influencing visibility earlier 

For years, product data mainly supported the ecommerce experience once the shopper arrived. It filled the product page, powered filters and helped customers narrow options across a retailer’s catalogue. 

 

Conversational discovery changes some of that flow. Customers can describe what they need in more natural language, refine the request and compare options without relying entirely on navigation menus or category structures. The product record starts doing more work before the product page is ever seen. 

 

A dress with unclear delivery information becomes harder to surface when the customer needs it for a specific date. A sofa with incomplete dimensions becomes less useful when the search is shaped around room size or access. A beauty product with limited suitability information becomes more difficult to match against a specific concern. 

 

The products themselves may still be commercially strong. They may be well priced, available and strategically important for the retailer. The surrounding information still needs to make that relevance clear enough for discovery systems to interpret. 

 

Retailers already hold much of this information, although it often sits across separate systems. Catalogue content, pricing, inventory and fulfillment data may be updated by different teams at different speeds.  

 

In a more traditional ecommerce journey, retailers had more opportunities to recover from some of those gaps through onsite merchandising, recommendations or the product page itself. Earlier comparison leaves less room for recovery because the shopper may never reach the page. 

 

 

Merchandising is moving closer to the product record 

Retailers have traditionally shaped discovery inside environments they controlled. Search ranking, promotional placement and merchandising all helped influence which products customers saw first and how they were framed against alternatives. 

 

AI-led discovery pushes part of that process into environments where retailers have less direct control over the comparison itself. A retailer may want to prioritise products with better availability, stronger margins or seasonal relevance, although those priorities only carry through if the underlying product information reflects them clearly enough. 

 

That starts to change the role of merchandising. Commercial priorities increasingly need to sit within the product information itself instead of being layered on later through onsite ranking rules or campaign planning. 

 

The effect is easiest to see in categories where customers compare carefully before buying. Fashion purchases depend heavily on fit, material and delivery timing, while homeware purchases depend on dimensions and fulfilment options. These details are now imperative in shaping how products are surfaced and compared. 

 

 

Accuracy becomes more commercially important 

Product accuracy and recency is often treated as an operational issue, although customers experience it much more directly. Wrong stock information damages trust quickly. Unclear delivery dates push shoppers elsewhere. Thin sizing or compatibility information makes comparison harder at the exact point customers are trying to narrow the field. 

 

Conversational discovery increases the pressure because these details are being used earlier in the buying journey. Retailers responding well to this are usually focusing less on producing dramatically longer descriptions and more on improving structure, ownership and consistency across the product record in more real time.

 

For many retailers, the practical work starts with understanding which attributes influence discovery most heavily in each category, then improving how those details move quickly and consistently across catalogue, pricing, inventory and fulfilment systems. A unified and highly accessible layer for data retrieval is critical.

 

The retailers best placed for this shift will treat product information as part of revenue performance. In a market where customers are comparing more carefully and margins are harder to protect, incomplete or fragmented product data is becoming increasingly expensive to ignore. 

 

---

 

Piyush Patel is Chief Ecosystem Officer at Algolia 

 

Main image courtesy of iStockPhoto.com and Vertigo3d