Skip to content 
By 2026, AI is expected to directly influence up to 20% of purchasing decisions in consumer goods. And increasingly, AI is not just the means of discovery. It is the buyer. Introducing the AI shopper.
Something is shifting in the way people buy everyday products, and most of the industry has not caught up with it yet.
The shift is this: AI is no longer just helping consumers find products. It is starting to make purchasing decisions on their behalf. Choosing brands. Switching retailers. Restocking households. In some categories, the consumer states what they need, and the AI does the rest. In others, it is recommending rather than buying outright. But the direction is clear, and it is accelerating.
For an industry that has spent decades perfecting the art of influencing human beings through advertising, media, and along the purchase journey, this is not a small adjustment. It is a different game entirely.
The machine does not care about your advertising
The traditional FMCG model is a push model. Manufacture the product, build the brand through advertising, negotiate shelf space, run promotions, and hope the consumer picks yours over the one next to it. The entire system is based around the premise that human beings can be influenced by brands through media and along the purchase process.
An AI agent is not influenced by media. It is scanning product metadata, ingredient lists, accreditations, nutritional data, and price. It is matching those facts against a specific consumer need. And it is doing this across every retailer simultaneously.
What we are probably seeing is the emergence of a pull model. The consumer states a need. The AI finds the product that best fits that need, wherever it sits, at whatever price. Shelf position is becoming increasingly less relevant. Advertising recall matters less to the machine. The checkout conversion funnel that retailers have spent years optimising matters less and less, because the AI is not browsing. It is buying.
The great squeeze
This creates a polarisation that will reshape category dynamics, though not uniformly. The pace and severity will vary by category, by market, and by how quickly consumers in each segment delegate decisions to AI.
At one end, AI drives commoditisation. For products consumers do not care about choosing, the ones they know they need but are not invested in selecting, the AI will find the cheapest option from wherever stocks it. Laundry detergent, toilet paper – if you are the number one brand in those categories, you are probably fine, because consumers will search for you by name. If you are the number two or three, your traditional tools for cutting through, disruptive advertising, off-shelf display, the promotional offer, the aisle-end promotion, are becoming less effective by the month.
At the other end, AI enables specificity. When consumers search for something that fits a particular need, a low-UPF breakfast cereal, the right food for a three-year-old Labrador, an energy drink that genuinely works for endurance athletes, the AI looks for the product whose factual profile best matches that requirement. Not the best-advertised one. The best one.
That is good news for brands built around genuine product excellence. It is very bad news for brands that occupy the middle ground: not famous enough to be searched by name, not specific enough to win on product-need fit.
The retailer has a new weapon. And a new problem.
Retailers are not passive in this. They control the AI layer on their own platforms. Tesco, Ocado, Sainsbury’s, they all have recommendation engines that steer consumers towards particular products. And unlike shelf placement, which a shopper can physically see and navigate around, an AI recommendation is invisible in its construction.
That gives retailers real power. They can default to own-label where margins are higher. They can prioritise suppliers who pay for prominence. They can steer consumers away from brands negotiating hard on wholesale price.
But there is a tension. Consumers are willing to delegate purchasing decisions to AI precisely because they trust it to be impartial. If it emerges that retailer AI systems are systematically biased towards commercial outcomes, that trust collapses. And it does not just collapse for one retailer. It takes the whole proposition down with it.
What we are seeing is a trust paradox: the more powerful the AI layer becomes, the more fragile the trust that sustains it.
What this means on Monday morning
If you are a Head of Insights or a marketing director at an FMCG company, the first thing to understand is where your category sits. Is it heading towards AI-driven commoditisation, where the fight is about brand fame and value? Or is it a space where consumers are actively searching for need-specific recommendations, where the fight is about product-need fit and machine-readable data?
The answer determines everything: how you develop products, how you describe them, how you track performance, and what you measure.
Because the AI is not reading your advertising. It is reading your ingredients list.
About the author
- Peter Kneale
- Global Head of Consumer, STRAT7
Claire is a Director in the consumer team, specialising in large scale, quantitative work designed to identify and unlock opportunities for brands as the FMCG landscape evolves at pace. She has extensive experience working for major beverage brands, food manufacturers and within consumer health. Her recent work focuses on how food and drink brands can continue to drive relevance as competitive sets extend beyond traditional boundaries.