Transforming the customer experience with AI

There are four key ways retailers can make it quicker and easier for shoppers to find what they want and accelerate the path to purchase using AI:

Smarter search

Retailers are yet to embrace AI in search fully, which means a significant opportunity remains untapped for early adopters to set the pace. Think of AI as being able to deliver a conversational experience for customers searching for items online.

For example, Mindera recently created a proof of concept where a shopper looking for a blue party dress could ask to add further specifications to narrow down their search. The shopper could say “not V-neck” to which AI would respond by removing all blue dresses with low cut necklines, without the need for the shopper finding and clicking a filter menu. This kind of smart search is going to become increasingly prevalent, in part driven by the fact that more and more shoppers will demand this level of functionality.

More sophisticated product chat

Retailers like Zalando are enabling customers to make more detailed searches through an AI-driven chatbot where shoppers can specify product attributes that keyword searching can’t handle - for example, specifying a t-shirt without brand logos or a specific style without knowing the technical term for it.

Ralph Lauren has a similar option integrated into its app.though it is currently separated from the brand’s ecommerce system, so shoppers have to move between the chatbot and the website. But it’s likely to be more embedded and therefore easier to use in the future.

ChatGPT also recently released a demo that enables users to search for products from retailers through the platform, specifying criteria like “gifts under £100”. Recommendations are then presented and purchases can be made directly on ChatGPT using agentic commerce protocol.

According to OpenAI, Shopify and Etsy already sell through the platform however the success of the programme remains to be seen. Meta tried something similar with in-app checkout on Facebook and eventually killed it - not because the technology failed, but because post-purchase complexity (returns, queries, fulfilment issues) is hard to own when you're the intermediary. That problem doesn't disappear with a different platform.

Deeper personalisation

Many retailers already use personalisation to improve the customer experience, and AI helps do this more effectively. However, this doesn’t work for businesses with low average transactions per customer. A lot of fashion businesses, for example, experience around 1.6 transactions a year per customer on average. Such minimal purchase and search data does not enable strong personalisation.

It generally works better in grocery businesses with around 50 or more products being purchased by individuals on average each month. With this amount of data, you can gauge customers’ habits quickly and understand what they're interested in.

Good personalisation relies on clean, unified, accessible and current data. As part of Mindera’s work on the Selfridges Unlocked loyalty programme, AI was used to clean up the iconic retailer’s customer database.

Any business that has physical stores and an ecommerce website like Selfridges will inevitably generate duplicate records. This can be down to human error from a shopper keying in their name incorrectly, or someone changing their name after getting married. Some businesses worsen the problem because they make unique IDs based on unusual things.

Cleansing data can be a tough job as there can be hundreds of thousands of different records. AI can accelerate the process if prompted in the right way. It’s ideal for big data sets, can score duplication probability, and works well with poorly structured data.

Find out more about retail tech trends in our upcoming Spring 2026 report.