How AI is reshaping product discovery in F&B

For decades, food and beverage companies vied for consumer attention on two fronts: the physical shelf and the digital shelf. But in the age of artificial intelligence, a ‘third shelf’ is emerging – and it is already redefining the way consumers discover and purchase products.

The third shelf is shaped by AI-driven agents and is no longer a hypothetical disruption to the industry. “It’s happening right here, right now,” said Hannah Law of market intelligence agency SPINS during a recent webinar.

What’s different about the ‘third shelf’ is how consumers interact with it. No longer are they scrolling down web pages and sifting through online catalogues to find products; instead, AI fetches the options straight to shoppers, complete with recipe and dietary recommendations.

In short: if finding the right dinner and drinks pairing ever felt effortful, agentic AI is narrowing those options and making choices quicker.   

Jessie Wright, senior vice-president at SPINS Foundry, said: “We’ve shifted away from keyword searching in a search bar to users asking questions like, ‘What’s the best magnesium for sleep?’ and getting a short list of responses, not a page of results.

“These AI agents are essentially curating recommendations for individuals, rather than presenting a catalogue they scroll through.

“If you’re not optimised to show up on this third shelf, you’re really losing the opportunity to surface to a user – and you might ultimately be invisible.”

So how big is the shift already?

According to MikMak, which tracks around $3bn in global commerce media spend, there has been a 30-fold increase in ChatGPT referral traffic to brand websites. This is a significant change in search patterns, and one that suggests consumers are already actively embracing generative AI not just to research, but as a purchasing tool, too.

This means brands must take their product pages AI-friendly – and that’s a different task altogether from traditional branding and marketing strategies. “A human responds to marketing copy,” said Wright. “An agent responds to structured data. If your data isn’t structured properly, you might have a data problem contributing to your visibility on this third shelf.”

In other words, pretty imagery and clever advertising would only get brands this far. In the AI era, they need to make product attributes ‘machine-readable’ – with ingredients, allergens, certifications, nutrition data and label claims all available in a clear, structured way so AI can pick it up and fetch it to the ‘third shelf’. Not only that: because generative AI ‘scrapes’ information from across formats and mediums, product makers must lean heavily on standardising their messaging across all levels of branding and marketing.

“A nicely-formatted product detail page with beautiful marketing copy can actually be nearly invisible to one of these AI agents if it’s not structured in the right way,” said Wright. “These agents reason over information in a very different way than traditional search ever did.”

So how can brands meet the needs of AI agents that aren’t wowed by crafty advertising? Through a multi-faceted approach that builds on existing expertise, rather than by starting from scratch.

“The first step is evaluating whether your product data is actually ready for agentic discovery,” Wright explained. “Brands should be thinking about whether their attributes are clear, consistent and available everywhere an agent might look for information about their products. The window to get ahead of this is right now.

“Brands that are actively using these AI tools can start to understand how demand might be shifting earlier, especially when it’s combined with velocity data and consumer research,” Wright added. “AI can be another input signal for innovation; not a replacement, but a powerful addition.”