Retail’s AI conversation is starting to feel oddly familiar. Consumers and retailers are moving at different speeds. This time, though, it’s the inverse of what typically happens. In the case of AI, consumers are demonstrating a readiness for technology that retailers haven’t been able to match.

That’s the core tension running through an extremely insightful new piece of research conducted by global AI and tech acceleration partner CI&T, called the ‘Retail Tech Reality Check’. Built on data from 2,000 consumers across the UK and Ireland, it makes for a read that may well challenge many of your preconceived ideas about the technology currently fuelling the retail industry.

Retail Gazette got the chance to sit down with Melissa Minkow, Global Director, Retail Strategy & Insights at CI&T, to unpack what the findings actually mean in practice, beyond the headlines, and where retailers are perhaps getting in their own way.

“It’s been a pivotal year,” she says. “I know we’re in unprecedented times over and over and over again, but it has been a really pivotal year from an economic perspective for consumers.” She points to changing spending behaviour, shifting promotional rhythms, and a UK consumer who is increasingly comfortable adopting new behaviours.

“We’re kind of at a tipping point, and AI is fuelling it. 64 per cent of consumers say that they think retailers should use AI to improve the shopping experience. This is six percentage points higher than what US consumers said,” she notes. “That alone was surprising. I didn’t really expect the warmer welcome over here!”

That ‘warmer welcome’ challenges a narrative that’s crept into retail’s AI discourse that the tech potential exists, but consumers are sceptical. In reality, consumer scepticism is more nuanced than that. People can be worried about privacy and still want convenience. They can distrust a retailer’s data handling and still use ChatGPT to shortcut research. They can be cautious and curious at the same time which, frankly, describes the modern shopper more accurately than any single label does.

 “And 61 per cent say that they have already used AI tools while shopping,” Minkow says, pointing to another revelation that hit hard. From her perspective, this tracks with a UK market that has often been faster to adopt digital commerce behaviours in certain categories. “I’ve repeatedly seen that there is faster ecommerce adoption here, like grocery,” she said, contrasting it with the US, where ecommerce grocery still lags behind expectations. The implication is that if the UK has already normalised certain digital habits, it makes sense that AI-assisted shopping will slip into everyday behaviour faster too.

And then there’s the wider mood. Minkow describes seeing the numbers around spending mindsets as revealing, even if the story is broadly familiar. “It was really high percentages of consumers just saying they’re pulling back on spend, and anticipating prices going up,” she notes. The difference with data, of course, is that it removes your ability to downplay what you’d rather not deal with. When the majority of shoppers tell you they’re budgeting harder, you can’t pretend loyalty will carry you through on relationship alone.

Yet some retailers simply don’t have the internal muscle yet. “They may not have the resources, they may not have the workforce that totally understands all of its capabilities,” Minkow says. Some are wary of early adoption because they’ve watched hype cycles play out badly in the past. “Being an early adopter in retail is worrying sometimes,” she adds. “We saw what happened with the Metaverse. Sometimes playing it safe can really pay off.”

But she draws a crucial distinction. AI looks less like a trend to her, and more like a competitive advantage that will compound. “I’ve even been surprised at my own excitement around AI and not thinking about it as a trend, but as a really crucial competitive advantage for retailers,” she says. That’s the point retailers need to sit with. Metaverse was a bet. AI is increasingly becoming infrastructure.

Of course, for those looking to truly make meaningful change in their organisation, the data has to be present and accurate. “Data is the foundation for AI,” Minkow says, “and you’d be surprised how many retailers don’t have their data in a good place to fully leverage AI. It’s not clean, it’s not streamlined, it’s very siloed.”

This data problem explains why so many AI initiatives get stuck at pilot stage. If your product data is inconsistent, your customer data fragmented, and your inventory logic unreliable, AI amplifies it and even leads to potentially damaging and incorrect conclusions.

It’s also worth unpacking what consumers even mean when they say they want retailers to ‘use AI’. The phrase is so broad it’s almost useless, like digital transformation or innovation, until you pin it down to a job to be done. For Minkow, the consumer definition of AI is remarkably simple. Make shopping easier and take away unwanted friction.

“The ways that they want retailers to leverage it really come down to a convenient, quick, efficient, easy shopping experience,” she says. “Consumers are so mission oriented. They just need decision support and they need help feeling certain about that.”

Minkow also makes an important point about retail’s track record of building the wrong things. “Previously retail has been kind of proactive in the wrong directions,” she says, “where you’ll see retailers offering capabilities that the consumer doesn’t care about.” If you’re a retailer and your AI strategy starts with ‘what can we build?’ rather than ‘what frustrates customers today?’, you’re already drifting.

Of course, privacy is still a headline concern. Especially in a UK retail market where high-profile incidents have made data security feel personal. But Minkow flags something that retailers should find even more unsettling, the fear that AI will be used to manipulate.

“Privacy is a huge concern,” she said, “but the secondary most common concern was this idea of being upsold or cross sold and spending too much, or the AI being biased.”

If AI recommendations feel like they’re engineered to push volume rather than help the customer, people will clock it, and they’ll resent it. Minkow’s advice here is direct: “Retailers really need to be transparent. Explain that this is truly intended to help you maximise your spend. It’s not about getting you to spend more, it’s about getting you to spend smarter.”

This is where the ethics conversation becomes commercially relevant. Bias doesn’t need to be malicious to be damaging. All it needs to be is noticeable. If the shopper feels nudged, the system stops feeling helpful and starts feeling suspicious.

The report also explores agentic AI, and it’s one of the areas where the consumer data is more advanced than the industry expects. Minkow is clear that most shoppers don’t know the term. “62 per cent had not heard the term agentic AI before,” she says, so CI&T defined it plainly. An assistant that can order items based on your preferences and purchase history.

What happened next is the interesting part: “58 per cent were interested once they read our definition,” she says, and the strongest interest was in routine purchases. The items people would happily put on autopilot. That’s where agentic AI becomes useful, in the weekly admin of life.

And that ties to one of Minkow’s most pointed observations. For many people, shopping isn’t a hobby, it’s just a task. “For a lot of consumers right now, retail is admin. Shopping is admin,” she says. That’s a brutal sentence if you’re a brand that has spent the last decade talking about experience as if everyone wants to be delighted every time they buy toothpaste.

She also challenged a belief that still sits deep in ecommerce thinking, that more time on site is always good. “A lot of times retailers have thought that the amount of time a consumer spends on your site is like the higher the time, the better the experience,” she notes. “And what I’ve tried to say is no, the faster the experience, the shorter the experience, the higher the NPS.” In other words, sometimes engagement is just trapped customers.

On physical retail, the report pushes back on another assumption that stores are fading into irrelevance. They’re not, they’re being reassigned. Minkow wasn’t surprised that shoppers still want to start journeys in-store across key categories because the reasons are human. Proximity, routine, and touch-and-feel certainty. However, agentic AI can still permeate and augment this experience to make it more efficient.

“I’m ready to start seeing pick up areas that say, ‘agent bought online, pick up in store’,” she laughs. “It’s a little different than buy online, pick up in store. We are going to have to go into brick and mortar and represent ourselves after an agent has represented us.”

As the conversation wraps, Minkow is keen to leave readers with some vital advice about what they can do today, to start aligning with the kind of future she’s discussed.

“Data is the foundation,” she says. “So many retailers are taking manual approaches. They’re still in literal spreadsheets. The faster you can get your data in a position to be easily automated, streamlined, secure, it will benefit your whole business no matter what. It literally benefits every single department. Get your data ducks in a row!”

If you want AI to be more than a nice pilot and a few internal experiments, it’s the only place to start.

To dig into the full findings, including the deeper breakdowns of consumer behaviours, concerns, and expectations, read CI&T’s full ‘Retail Tech Reality Check’ report.

Click here to sign up to Retail Gazette‘s free daily email newsletter