How to regain control of the customer experience

For the past few years, much of the conversation around AI and commerce has focused on discovery: how people search, how they ask questions, and how they've shifted from typing keywords into Google to holding conversations with tools like ChatGPT and Gemini.

But shoppers aren’t just asking AI for recommendations. They increasingly expect AI to complete tasks on their behalf.

Autonomous buying sounds futuristic until you look at how quickly behaviour has changed over the last two years alone. Moving from product recommendation to action is a far smaller leap than many businesses realise.

Today’s buyers might ask an assistant to purchase a gift within a specific budget, or organise travel based on a set of requirements. The customer still defines the requirements, but increasingly the software is handling the work needed to get the job done.

The takeaway? AI is no longer acting merely as a source of information; it’s becoming an active participant in the buying journey.

For time-poor consumers, outsourcing decisions to a chatbot is a win. For retailers, watching someone else play matchmaker with your customers is less fun.

Losing the Moment

So what does it actually mean for retailers when a customer turns to a third-party AI?

It means you lose the moment.

You no longer get to make personalised recommendations or guide customers to the right product. That relationship now sits with the AI, not with your brand.

Data is at the heart of everything you do. When a shopper uses your search bar, you have insight into what they're looking at and how often. When they turn to a third-party AI assistant, that first-party data is lost, and it's the AI that builds a real understanding of who's shopping and why, not you.

The problems don't stop there. ChatGPT doesn't ask follow-up questions, or upsell. It won't notice that someone browsing for a coat might also need boots, or nudge them towards the better-fitting option. Worse still, it might point a customer towards a competitor, or serve up incorrect product information you have no control over.

Taking Back Control

So how do you win back control?

You give shoppers an on-site experience that rivals general AI.

Shoppers should be able to ask your website’s search bar for what they want, in plain language, and get the right answer straight away, without wading through hundreds of options.

On a fashion site,

"a light yellow dress for a summer wedding"

should bring back exactly that. Not a dead end because nothing matched the keywords, and not thousands of "dresses" because that was the only word the search picked up on.

Putting It Into Practice

Shopping center owner Hammerson's shoppers weren't browsing for inspiration; they wanted to make fast, confident decisions about where to go and what to prioritise on their trip. But slow load times, dense layouts and limited discovery tools were getting in the way at the very moment they were choosing which brands to visit.

Firney redesigned the on-site search experience across several of Hammerson's major shopping centre websites. The old search demanded exact brand knowledge, struggled with spelling variations and often led to dead ends, limiting engagement and exploration alike.

In its place, Firney built a real-time search with intelligent auto-suggest that links retailers, categories and related brands, actively encouraging discovery rather than abandonment.

The next step is natural language processing, so that the search can interpret intent directly from conversational queries.

The impact is clear. Firney's study, The £4 Trillion Conversation, found that shoppers using AI-powered search convert at rates 200–300 per cent higher than those relying on standard filters and category pages.

Search is Only the Start

Once you understand what a shopper is asking for, the job is to guide them the way a good in-store assistant would: offering products that fit the problem, upselling and cross-selling where it makes sense, and factoring in what's actually in stock.

A customer searching for a "smart suit for a wedding" should be met with a confident recommendation, a reason it's the right one, and the shirt and shoes to go with it.

It's the shopkeeper experience, rebuilt for a smartphone or web-based interaction.

Do it for the data

Every conversation a shopper has on your site generates proprietary intent data: what they want, how they ask for it, and what finally makes them buy.

Feed that intel back into your products, content and platform, and you build an AI experience powered by your own data. One third-party AI can never replicate because it never sees it.