Is the shopping list about to be consigned to the dustbin of history? Yes it is. Not because people will stop buying things, but because they may no longer need to decide which things to buy, says German Faraoni Heidenreich, global director, data & shared services at Reckitt, ahead of his panel at CustomerX.
Consider the familiar chaos of preparing for a busy weekend. The house needs cleaning, the fridge needs stocking and the children need taking to sports practice – or in my case this weekend, university. Today, that means remembering what is required, searching for products, comparing prices, choosing retailers and arranging collection or delivery.
But, by the time my next child heads to uni in two years’ time, I may simply tell an AI agent what the weekend involves and ask it to sort it all out. I for one am keen.
This is a far more profound change than merely using AI to make the existing shopping journey faster. It potentially replaces much of that journey with something based not on products, brands or retailers, but on outcomes.
German Faraoni Heidenreich, Global Director, data & shared services at Reckitt and a speaker at the forthcoming CustomerX event, believes this shift is already starting to reshape how consumer goods companies think about commerce.
“People in the past would say, ‘I know I have to go and clean. I have to go and check if this type of retailer of my choice will have more or less the brand that I heard about, or a friend told me about, or my parents have been using,’” he says. “Right now, a key shift is that they’re actually thinking about the outcome that they want. ‘This weekend I need to clean the house and get the kids to football,’ and that is very different.”
Outcomes, not sectors
This is very much a CustomerX view of the world. Customers do not experience their lives as a collection of sectors. They do not necessarily think in terms of grocery, household care, transport, entertainment and sport. They think, in Heidenreich’s example, about getting the family through the weekend.
As a result, an AI agent could become the mechanism that joins these currently disconnected needs together. It may understand the customer’s budget, preferred brands, sustainability expectations, location and schedule. It could then decide what is required, where items should come from and how these should be delivered.
“The 2030 version of commerce will be agentic and it will be humans articulating outcomes that they want to reach,” says Heidenreich.
The implications for brands are considerable. A customer may no longer search for a particular product, encounter a conventional advert or consciously choose a retailer. Instead, an agent could assess price, availability, suitability and delivery before presenting a recommendation – or eventually completing the purchase itself.
For Reckitt, that creates a new strategic question. “How do we show up in a relevant way for that AI agent that might be making a choice into: should I buy this product versus this one?” Heidenreich asks.
Brands, in other words, are acquiring a second audience. They must continue to earn the attention and confidence of people, but they must also become visible, understandable and relevant to the machines acting on their behalf.
The retailer becomes blurry
Consumer goods companies have traditionally used advertising to generate demand and then worked with retailers to make sure products are available when that demand arrives. Agentic commerce makes that relatively straightforward route more complicated.
“We needed to create the proper buzz with consumers so that they would be driven into the retailers,” says Heidenreich. “Now, it’s not about some particular retailer. We need to speak to an ecosystem because that consumer, based on their context, may not depend anymore on a particular type of retailer.”
An agent might choose a supermarket, marketplace, delivery app or personal-shopping service according to what is most appropriate at that particular moment. As Heidenreich puts it: “The retailer starts becoming a blurry thing”.
Retailers do not disappear from this picture. Products still have to be stocked, sold and delivered by someone. What weakens is the assumption that the customer begins with a retailer and completes the entire journey inside its ecosystem.
“We stop having that kind of control,” he says. “Does it make it more difficult? I wouldn’t say so. Does it make it more complex? Yes.”
Responding to that complexity requires much tighter connections between brands, retailers, platforms and other partners. It also requires data to move more fluidly between them. “If it doesn’t happen in that way,” he warns, “then we miss that sale.”
Selling to the machine
This is not solely a prediction about 2030 – the theme of Heidenreich’s CustomerX Think Tank session. Reckitt is already changing how it presents information about its products so that algorithms can understand them more completely.
Online product descriptions have traditionally been written for people: concise, persuasive and relatively easy to scan. Machines do not share those preferences. They can process far more information in an instant and use additional data points to determine whether a product fits a customer’s requirements.
“What [Reckitt] has been doing is putting much more information out there that is machine-readable,” Heidenreich explains. “In many cases, we tripled the content for machines. It’s not human-readable; it’s machine-readable.”
The company has already seen an uplift in algorithms presenting its products to consumers as a result.
This is a significant change in how brands think about discoverability. Search optimisation has largely been about interpreting the words typed by humans and securing a prominent position in the resulting list. Agentic discovery is more likely to depend on giving machines enough structured, accurate information to judge whether a product provides the best answer to a specific need.
“We are no longer going to fight for that ad attention or try to present ourselves as the best object to that human,” says Heidenreich. “It is going to be: how do we better address that outcome?”
Advertising money will follow. He expects spending to be redirected from familiar media channels towards the algorithms and platforms influencing agentic decisions. “We will have to start paying media money, instead of to a media outlet, to somebody that created algorithms,” he predicts.
Technology is not the only obstacle
But there are challenges. For example, the biggest barrier may not be the capability of AI, but the ability of businesses to organise themselves around it.
“Are our marketing plans more connected with the technical ability to enrich all this information for the agents to actually find us?” Heidenreich asks. “Are all our internal data points clear enough so that, when we expose them, they make sense out there – or did we just show a lot of dirty laundry?”
Marketing teams have become adept at testing campaigns and learning quickly. Connecting those experiments to product data, technical infrastructure and external partners is often much slower. Businesses do not need a perfect system before beginning, but they do need what Heidenreich calls the “bare minimum” necessary to start learning.
They must also build privacy, governance and cybersecurity into the model from the outset. The more information and connections an organisation exposes to external systems, the greater the need to ensure that those connections are secure and compliant.
No company can resolve all of this alone – and no single sector possesses all the answers. This is what Heidenreich hopes to explore at CustomerX: whether other businesses are seeing the same signals, how those furthest ahead are behaving and, crucially, whom they are choosing as partners.
“CustomerX provides the perfect setting for those kinds of dialogue,” he says.
It is an appropriate ambition for an event built around customers rather than sectors. If people increasingly ask commerce to deliver outcomes that stretch across traditional industry boundaries, the businesses hoping to provide those outcomes will have to cross those boundaries too.
To hear German Faraoni Heidenreich, global director, data & shared services at Reckitt, discuss the business landscape in 2030, register for CustomerX. This interview first appeared on the CustomerX Substack.
