Can food brands survive the rise of agentic commerce?

Agentic commerce, where AI agents act on behalf of consumers, is here and it’s moving fast. Some brands will win; others will lose.

Food industry advisory and consultancy business Bramble Intelligence recently released its Tomorrow’s Appetites report, which delved into the three forces that will reshape food and beverage, one of which is agentic commerce. Here, managing partner Chris Mitchell unpacks what it will mean for manufacturers.

What will agentic commerce look like once it’s fully developed?

There will be two versions, says Mitchell. “Retailers will have their own agents as they’ll want to maintain their customer relationship by keeping them within their environment. It’s basically becoming a food concierge for you. It’s asking how many people are in your household? What flavours do they like? What intolerances or allergies do they have?”

This prediction indicates that agentic commerce won’t stop people going to bricks and mortar stores. “The economics of it are that until we have autonomous driving and home delivery is genuinely more efficient in terms of both time and cost, people will still go to grocery stores,” he says.

The other version of agentic commerce is the agent that sits outside the retailer and it’s going to be a native AI application, probably from one of the leading models such as OpenAI, Anthropic or Google. “What that’s going to be doing is aggregating your split shop across all the retailers, making sure that you’re getting the best price and the best nutrition advice,” says Mitchell. “It will be advocating on your behalf.”

Will brands and brand loyalty still matter?

“In one way, brands won’t matter,” says Mitchell. “AI doesn’t care what kind of cartoon character you put on the pack. It’s just looking at the facts.” This is where the importance of companies working out how to make information on their products machine readable – data formatted so that AI can process and understand it without human intervention – comes in.

But brands won’t lose their power altogether. “When you find one that matches your own definition of nutritional performance and where you have an incredibly high degree of trust, we talk about this as the second coming of the brand. Brands started many years ago because food was killing people. And, literally, people put their names on it; purveyors of food said, trust me.”

Now that’s happening again as some people look to regain trust in Big Food, but companies won’t be able to rely on heritage alone.

“Those brands will have to stand up on food quality,” says Mitchell, “but we will definitely continue to have them because they help you to quickly figure out if they are aligned to your values, to your purchase criteria, and ultimately to your identity. The highest form of a brand is that it’s part of who you feel you are. It represents you as a person.”

When AI and the consumer align on that choice, it hits a sweet spot. “This is where you’re going to have a bit of a Venn diagram,” says Mitchell. “When the consumer and the AI agree, the brand strength grows.”

The problems emerge when the two disagree. “By comparing products, AI is forcing shoppers to look at them in a new light. As soon as the AI is saying, hey, you should buy X and not Y, and the reasons are beyond just the price, consumers start to say, oh, I hadn’t thought about that.”

The Tomorrow’s Appetites report said that the companies that will succeed will be the ones that remain closest to the customer. How can they do this when AI begins to stand between them?

In an old-school way, direct contact with consumers will be key.

“We’re in discussions with clients around their innovation agenda and what’s been interesting is to hear how much they are going out to consumers directly,” says Mitchell. “Yes, AI can help them discover ingredients or new formulations, but in order to put significant capital behind a new product or formulation, they’re going to have to go out to consumers. It has always been part of the process, but that’s how brands are going to be able to keep that connection.”

It’s also likely that we’ll see more pop-ups with brands, again, trying to establish a direct connection to consumers. Ultimately, “the brands that really understand their customers best will win in their categories”.

Which categories do you think will be most vulnerable to agentic commerce?

Ultra-processed foods are likely to struggle, according to Mitchell. “Those that are more composite, more reliant on additives and preservatives, those may be tricky,” he predicts. “So breakfast cereals, snacking products, some ready meals. And then I think because of GLP-1s, you have a compound problem. If you have a sweet bakery product, that’s served in larger formats, full of additives, that’s going to have a harder time because people want smaller portion sizes, cleaner labels and products that provide more functional benefits.”

But it’s not all trouble, according to Mitchell. Categories that were historically just seen as commodities may do well. “I think we’re going to start to see more differentiation, maybe even more branding entering into commodity categories - meat, fish and poultry, fresh fruits and vegetables, dairy.”

As people shift out of some of the ultra-processed categories and buy fresher food and proteins, they are starting to ask AI more questions about those foods, such as omega 6/omega 3 ratios or provenance, so brands can then differentiate themselves. “We’ve habituated people to be more inquisitive about their food generally, and now they’re extending that into categories that people thought were just commodities,” says Mitchell.

What do brands need to be doing now to gain the trust of an AI agent?

Initially, brands must create a broad and reliable digital footprint to feed the AI. “Brands will reach AI directly by having a lot of published content,” says Mitchell. “You need to be appearing in trade magazines, and appearing in top 50 or top 10 lists in the category. I think things like the Great Taste Awards will be important. AI is currently scraping, so it could be Substacks, it could be traditional mainstream media, it could be bloggers or influencers.

“And it’s looking at four criteria: the sensory experience at the time of consumption, the nutritional performance, the avoidance of toxins and the provenance, so brands need to be putting stuff out across all the dimensions of quality into as many different forms as possible.”

However, reliance on this type of information gathering may not last. “This works today but it’s a transitional phase,” says Mitchell. AI companies are investing in systems that attempt to give food and drinks a definitive score, at least nutritionally.

That’s why brands need to do more than just produce information for the machines to read, “they need to think, how do we stack up on price and nutritional performance versus everybody else? At the end of the day, consumers and the machines need to agree on some absolutes,” concludes Mitchell.