How to run an assortment review with AI support

In merchandising, AI can automate data consolidation, forecasting, and promotional planning to unlock capacity gains of about 40 percent, according to McKinsey analysis published in January 2026. But which steps in a typical assortment review can AI truly accelerate, and which still demand the judgement of a category manager or buyer? This guide maps the process, drawing on recent practitioner cases and industry data.

What data can AI analyse before the assortment review?

AI-powered retail analytics platforms now consolidate data from point-of-sale systems, promotional calendars, weather feeds, and social media signals into a single view. Modern platforms use machine learning models trained on two or more years of historical sales data, promotional calendars, seasonality, and external factors such as weather, local events, and economic indicators. AI cleans sales data to remove the impact of stockouts and infers true demand, addressing one of the most persistent distortions in retail forecasting.

Fashion retailers that have implemented AI demand forecasting report reductions in forecast error of 30 to 50 percent compared to traditional methods. For a retailer with 20 million euros in annual inventory purchases, a 35 percent reduction in forecast error translates directly into millions in reduced excess inventory, reduced markdown losses, and reduced stockouts. At Zara, AI-powered demand forecasting now drives 85 percent of initial production allocation decisions, according to reports from January 2026.

Which forecasting tasks can platforms automate?

AI refines assortment planning by combining predictive analytics with product-level insights. It helps forecasts demand by style, colour, and region, highlights underperforming SKUs and identifies assortment gaps, automatically builds dynamic store clusters, and optimises size curves to reflect true demand. These capabilities help planners design assortments that match customer demand and minimise overbuying or missed opportunities.

Trend forecasting is another area where AI has demonstrated measurable speed gains. Zara is using AI trend detection tools that analyse social media, runway shows, and street style to identify emerging fashion trends three to four weeks faster than traditional methods, according to a March 2026 update. Zara's in-house AI platform identifies emerging trends three to four weeks faster than traditional methods, optimises assortment to better match local demand, and enables more precise allocation and replenishment, driving higher availability, full-price sell-through, and ultimately sales.

How does AI support the assortment review meeting?

During the review itself, AI surfaces recommendations rather than making autonomous decisions. Decision-making no longer needs to follow static weekly cycles, since AI agents can continuously tune assortments, adjust prices, refine promotions, and deliver rich, real-time insights. Target is using AI to rebuild its merchandising authority. Target chief operating officer Michael Fiddelke described how merchants can now instantly access trend insights. By leveraging AI to capture colour, material, style, and product details, and applying consumer research and brand principles, they can deliver unique and on-trend products to guests faster than ever before.

The role of the merchant is shifting from building the plan to editing what the model proposes. In 2026, AI replaces the analytical core of buying, including forecasting, assortment planning, and product discovery, but not the whole role. Human buyers retain negotiation, taste arbitration, and accountability for decisions. The realistic outcome is a hybrid: a model proposes the assortment and a human edits it.

Where does human judgement still matter most?

71 percent of merchants in a new McKinsey survey say that AI merchandising tools have had limited to no effect on their business so far. The challenge often lies less in the technology than in how it is integrated and used, according to the McKinsey Global Merchant Survey fielded in December 2025. This gap reflects the distance between algorithmic suggestion and commercial execution.

Merchants continue to own three critical tasks that AI cannot yet perform autonomously:

• Negotiating terms with suppliers and brands, where relationship context and strategic leverage matter
• Applying brand and aesthetic judgement to ensure product selections align with the retailer's identity and customer expectations
• Taking accountability for the final assortment decision, including exceptions and risks that fall outside the model's training data

A Walmart spokesperson told Modern Retail that when it comes to category planning and merchandising, it is personal, local and human. Applying AI is really to help serve customers and associates better so they can spend more time on the human aspects of their day-to-day lives instead of tactical, manual work.

Which vendors provide AI-driven assortment analytics?

The market for AI-driven assortment planning software includes both established enterprise platforms and newer specialist tools. Oracle Retail Analytics, part of Oracle Retail Insights Cloud Service, is built for large retail chains that need pre-built retail metrics without constructing a data model from scratch. It ships with 200 or more retail-specific metrics covering sell-through, stock cover, margin, and promotional uplift, and pairs with Oracle Retail AI Foundation for demand forecasting, markdown optimisation, and inventory allocation.

Blue Yonder is one of the most established AI-driven retail planning platforms, with deep capabilities in demand forecasting, replenishment, supply chain planning, and workforce management. Blue Yonder's Luminate platform uses machine learning to generate demand forecasts at SKU and store level, incorporating external signals including weather, promotions, local events, and economic indicators alongside internal sales history. Nextail is an AI-native merchandising platform built specifically for fashion and apparel retailers, focusing on inventory optimisation, allocation, and replenishment across store networks and e-commerce. Nextail's machine learning models account for the particular complexity of fashion retail, including short product lifecycles, high SKU proliferation, size and colour fragmentation, and the interaction between online and physical inventory, to optimise stock distribution and minimise both overstock and lost sales.

Platforms from RELEX Solutions, Impact Analytics, and o9 Solutions also appear in recent industry assessments, with each offering different strengths in demand sensing, scenario planning, or integration across the planning cycle.

What implementation patterns are retailers following?

Successful implementations tend to follow a staged approach. Retailers who implemented comprehensive systems reported a 32 percent average reduction in stockouts and a 28 percent decrease in excess inventory costs within the first year of implementation, according to case studies cited in a January 2025 analysis. The pattern typically begins with demand forecasting, then extends to assortment optimisation and allocation.

Only 24 percent of merchants in the McKinsey survey say that their organisations currently provide moderate-to-extensive AI upskilling. But they will also restore focus on the fundamentals: shaping strategy and building the assortment rather than generating reports and reacting to issues. Training is a common implementation gap, alongside data quality and integration challenges.

The most advanced retailers are testing AI-driven scenario simulation before committing to an assortment change. Target's AI approach extends to pretesting assortments using synthetic audiences, AI-driven models that simulate real consumer populations, according to recent remarks from Target's chief operating officer. This allows merchants to evaluate multiple options before the buy is locked.