How Nordstrom uses AI for personalisation, search and stylists

Nordstrom uses AI mainly in its shopping app, where generative AI helps produce trend and styling content alongside its stylists, personalised recommendations learn from each shopper's behaviour, and search has been rebuilt to be faster and more precise. Human stylists remain one tap away. The retailer has described these features publicly but has not released performance figures for them.

What AI features did Nordstrom add to its app?

In November 2024, Nordstrom unveiled a refreshed mobile app as the centrepiece of its holiday strategy, Retail Dive reported. Chief Marketer, citing Nordstrom's press release, said the redesign added more editorial content created with generative AI, better search, better personalisation and a larger role for the loyalty programme, with the aim of making shopping feel more inspiring and more personal.

How does Nordstrom combine stylists and generative AI?

Nordstrom's approach blends the two rather than replacing one with the other. Retail Dive reported that the app's trend reports combine the expertise of Nordstrom stylists with AI to surface relevant trends, and Chief Marketer described a section called Discover the Trend that sources content from stylists and generative AI.

This matters for a department store whose brand has long been associated with service. Generative AI lowers the cost of producing a steady flow of editorial content across many categories, while stylist involvement keeps that content anchored in the retailer's own point of view on fashion. It also reduces a known risk of generative content, generic or inaccurate output, by keeping experts in the loop.

Read also

How Kering uses AI: luxury clienteling and conversational commerce

How does Nordstrom personalise recommendations?

Two mechanisms were described. Style Swipes recommends products based on a shopper's habits and preferences, which suggests a simple interface for gathering explicit feedback. And Nordstrom's press release, quoted by Chief Marketer, states that the more customers interact with the app, the better it gets to know them, producing more personalised recommendations over time through enhanced learning features.

Personalisation of this kind depends on several data inputs, most of which a large omnichannel retailer already holds:

• Browsing and purchase history across app, web and, where linked, stores.
• Explicit preference signals, such as swipes, saved items and size choices.
• Loyalty programme data, which Nordstrom now shows prominently on the app homepage.
• Well-structured product attributes, so that recommendations match on style, fit and material rather than only on category.

How has Nordstrom improved search?

Nordstrom says search in the refreshed app is faster and more accurate, according to Chief Marketer. Shoppers can add a size, colour, price or material to the search box to refine results. For fashion retailers, this kind of attribute-aware search is often a quicker win than conversational interfaces, because it addresses the most common reason for abandoning a search: results that ignore what the shopper actually typed.

Search quality also depends on the catalogue behind it. A multi-brand retailer receives product data from many vendors, often with inconsistent naming for colours, materials and fits. Filters by size, colour, price or material only work when those attributes are captured consistently, so much of the effort behind better search typically sits in product data management rather than in the search algorithm itself. Chief Marketer also noted that Nordstrom had launched a marketplace in 2024 adding 300 new brands, which increases both the range on offer and the need for consistent product information.

Where do human stylists fit in a digital journey?

Retail Dive reported that customers can request a personal look from a stylist or book an in-store visit directly from the app. This is the clienteling side of the case: AI handles discovery and routine recommendations at scale, while the app routes high-intent or complex requests to people.

Nordstrom appears to be exploring a further step. A session listed in the Google Cloud Next 2026 event catalogue described Nordstrom using Gemini Enterprise for Customer Experience to build a shopping agent, presented as an AI stylist that uses multimodal AI to understand a shopper's style and intent. The listing also showed a cancelled status, and Nordstrom has not published details of such an agent, so it should be read as a direction of travel rather than a confirmed launch.

Read also

AI in clienteling: how next-best-message works, and where privacy sets limits

What can other retailers learn from Nordstrom?

The case offers a practical model for department stores and multi-brand retailers:

1. Use generative AI to scale editorial content, but keep stylists responsible for the point of view.
2. Collect explicit preference signals (such as swipes) alongside behavioural data to make personalisation more accurate and more transparent.
3. Fix attribute-based search before investing in conversational features.
4. Keep a visible route to human service, so AI raises rather than replaces the service promise.
5. Define success measures before launch and report them, since published outcomes are what distinguish a deployment from a feature list.

The limits are those of any personalisation programme: it depends on customer consent and data quality, it can narrow what shoppers see if recommendations are over-tuned, and its value is hard to judge from the outside without disclosed results.