Retail and Artificial Intelligence: Forecasting That Works, Pricing That Pays, and Shoppers Who Now Ask Chatbots

Abstract

Purpose. Retail was among the first industries to use machine learning on a large scale, in forecasting and pricing, and is now watching shoppers move part of their search to AI assistants. This article asks where artificial intelligence stands in retail in 2026 and where the industry may be by about 2035. Design and methods. A review combining the share of the retail research literature that involves AI (OpenAlex, 2010–2025), Eurostat's survey of AI use in EU retail trade enterprises (2021–2025), a forecasting competition run on retailer data, a field experiment in pricing, a study of generative AI in customer support, and published web-traffic statistics. Claims taken from the literature were checked against the sources cited. The scenarios are judgements built on this evidence. Findings. The share of retail works that mention AI rose from 3.4% in 2018 to 16.6% in 2025, and language-model works from 61 to 1,161. In the EU, 15.5% of retail trade enterprises used AI in 2025, up from 6.0% in 2021 and below the 19.9% for all enterprises, mostly text mining (8.2%) and language generation (6.0%), with 7.5% using it for marketing or sales. Evidence of benefit exists for forecasting and pricing: in a forecasting competition on Walmart data, the winning machine-learning submission beat the best statistical benchmark by 22.4% overall but only by 3.4% at the most granular level, and in a field experiment an online retailer's machine-learning pricing tool raised revenue of the test group by an estimated 9.7% (90% interval 2.3% to 17.8%). Generative AI in customer support raised resolutions per hour by 15% on average in one large study. A new development is that shoppers use AI assistants: traffic from generative AI sources to US retail sites rose by 693% in the 2025 holiday season, from a small base. We expect forecasting and personalisation to remain the core of retail AI through 2028, AI-mediated shopping to change how retailers are found and compared in 2029–2032, and the balance of power between retailers, platforms and AI assistants to stay open. Originality. The article ties together long-established retail uses with the new question of AI-mediated shopping, using adoption data for retail trade proper and results that show where the benefit shrinks as the task becomes more granular.