.png)
Picture this: A shopper opens ChatGPT and types, "I need a shampoo for curly hair that's vegan and won't irritate sensitive skin." Within seconds, they receive personalised recommendations. Not from ads or sponsored placements, but from an AI agent parsing thousands of product claims to find the perfect match.
This isn't a far-off scenario. It's happening now. While AI-powered shopping currently represents just 5-10% of online purchases, projections show this will surge to 20-30% by 2027 - a 400% growth rate that will fundamentally reshape how consumers discover and buy products.
In a recent webinar hosted by Provenance, founder Jessi Baker MBE convened three experts at the forefront of this transformation to explore what brands must do now: Brock Simon (Partner at Bain & Company, AI advisor to Kroger and Tesco), Giles Pavey (AI advisor to Unilever with three decades in retail data science), and Jennifer Blease-Williams (Digital Marketing Lead at sustainable beauty brand Faith in Nature). Their insights reveal not just where AI commerce is heading, but the practical steps brands can take today to ensure visibility tomorrow.
The question isn't whether AI will reshape how consumers discover and buy products, it's whether your brand will be visible when they search. Here's what early movers are learning, and why waiting could cost you the most valuable customers in retail.
The shift isn't just about AI answering product queries - it's about fundamentally changing where in the buying journey consumers engage with technology.
"To me, the bigger shift is watching that discovery move upstream," Simon explained. "You don't go, 'hey, this is the exact product I want, I'm gonna use AI to help me find where that is.' It's 'there is a family member I'm trying to buy for, there is a shopping list I'm trying to do, there's some problem I'm trying to solve that's really important to me.'"
Instead of searching for "organic shampoo," shoppers now describe complex needs: a family member's gift preferences, dietary restrictions, or sustainability priorities. The AI then narrows possibilities, finds products, and increasingly, facilitates the purchase itself.
A frustrating paradox has long plagued sustainable business: few consumers who report positive attitudes toward eco-friendly products and services follow through with their wallets. According to a Harvard Business Review study, 65% of consumers say they want to buy purpose-driven brands, yet only about 26% actually do so.
For Pavey, the most profound impact may be AI's ability to bridge what he calls the "think versus do gap."
"A lot of shoppers, when you talk to them about what they want to do and what's important to them, they'll say sustainability, they'll say provenance," Pavey noted. "But when it comes to being in a supermarket, their mind might be on other things—they might have young children with them, they might be time-pressured. More base, reptile brain versus primate brain takes over."
AI shopping assistants can keep consumers true to their stated values even when cognitive load would otherwise derail those intentions. By allowing shoppers to program preferences once - buy British, prioritise sustainability, avoid allergens - AI can filter thousands of products automatically.
Key predictions from the panel:
"Still today, there's not the systems yet in place to ground the difference between a semantic marketing claim and a true marketing fact," Simon warned. "You'll see a lot—if you ask for something to be organic, AI will make the leap and go, 'oh yes, organic, all 100% everything else.' And somewhere down that line, it went from a loose marketing claim to a fact that should have been verified and checked."
LLMs are prediction engines trained to generate plausible-sounding text - not to verify factual accuracy. When product data is fragmented or unstructured, AI fills gaps with likely-sounding content. A product that's 95% organic becomes 100% organic. A "natural" ingredient list becomes "all-natural."
For Faith in Nature, this manifested in frustrating ways:
"When we first started looking at this, we found quite often we weren't showing up in searches, or if we were, they were recommending the wrong product for the wrong hair type," Blease-Williams shared.
The solution requires two elements: structured data organised in machine-readable formats, and third-party verification that separates marketing language from substantiated facts.
Faith in Nature, a 50+-year-old UK-based sustainable personal care brand, offers a compelling case study in rapid AI commerce optimisation.
The brand had rich product claims (vegan formulations, sustainable ingredients, specific efficacy for hair types) but this data existed in fragmented formats. Their JavaScript-heavy website wasn't easily crawlable by AI. They were seeing just 0.3% of traffic from AI search, though notably, that small amount of traffic was converting at significantly higher rates than traditional channels, suggesting AI-sourced shoppers were highly qualified buyers.
Faith in Nature partnered with Provenance to implement structured "proof points" - validated product claims organised as machine-readable data with clear provenance and third-party verification. Rather than stating "vegan" in marketing copy alone, each claim includes the specific standard applied, verification source and date, structured schema markup, and an audit trail.
The test began with just 37 products out of 200+ on their site - a low-risk experiment that could scale if successful.
After just 4-6 weeks, the impact was measurable:
"I was shocked, to be honest, when you pulled them up," Blease-Williams admitted. "In my head, it is making that shift from SEO to AI, because obviously, the world of SEO these days, it takes a lot longer than that to see real-time changes. It makes it quite exciting that you can see such real changes in such a short time frame."
The universal lessons:
As Blease-Williams noted:
"AI does level the playing field a little bit. It's just data that you're putting out there, and not having to compete with other companies' marketing budgets."
Want to hear these insights directly from the experts? Watch the complete 45-minute webinar here
Webinar poll results revealed an industry-wide challenge:
As Pavey put it: "This is a one-way street. From what I see in large FMCG companies and working with large retailers, they are moving in an unstoppable way to AI doing their logistics, and that will inevitably mean they'll want suppliers' claims and know that they're valid."
Primary motivations for AI data optimisation:
Traditional SEO optimises for keyword density, backlinks, and page authority. AEO requires semantic clarity and factual grounding. It's not enough to repeat keywords. AI needs to understand what your product is and validate your claims.
"AI is thinking about semantics, what are the right words, and increasingly, they're building structure around it," Simon explained. "When to triage between marketing claims and being able to generate things, and when to ground in very specific facts."
Pavey offered an important starting point: "A good website for humans is a good website for SEOs, is a good website for AI." His recommendation: Double down on clarity. Make it unmistakably clear what is vegan. If something isn't vegan, ensure absence doesn't just mean "no data"—it means definitively not vegan.
Key differences:
Here's a reality that might surprise you: When a consumer asks an LLM for product recommendations, your brand website is not the preferred source. The preferred sources? Reddit. Review sites. Third-party testing organisations. Industry publications.
LLMs face the same challenge consumers do: distinguishing between marketing claims and verified facts. When a brand says "eco-friendly" or "dermatologist tested," should the AI trust that claim?
This is where third-party verification becomes critical. When an LLM encounters a "proof point" marked as verified by an independent authority like Provenance, it signals:
Why third-party verification matters:
Simon framed AI as a collaborative partner currently in its "norming stage."
"We are in the norming stage, both in terms of AI models learning how to chat with us in these new ways, and consumers learning what works best," Simon explained. "If you participate now, you're going to be part of that process of how we're shaping the thinking around everything from how we think about products to sustainability claims. If you get in later, you'll be trying to shift something already settled versus something that's forming right now."
The rules are being written now. Early participants help shape how AI interprets product categories and makes recommendations. Late entrants will adapt to norms they had no hand in creating.
Two-dimensional barcodes: The shift from traditional to 2D barcodes will make product information instantly accessible via smartphone scans - powered by the same structured data that feeds AI.
Retailer requirements: Major retailers are already incorporating AI into logistics. Claims data will increasingly become part of listing conversations. As Pavey noted: "If you don't do it now, somebody like Tesco is going to be demanding that you did it. And then it'll become a cost rather than an opportunity."
Why timing matters:
The transformation underway represents a fundamental rebalancing of power in retail. For decades, success correlated with marketing budgets. AI commerce changes the equation.
As Blease-Williams observed: "It does level the playing field a little bit. It's just data that you're putting out there, and not having to compete with other companies' marketing budgets."
When an AI agent evaluates which product best matches user needs, it doesn't care about ad spend. It cares about structured, verified facts. The brand with the clearest data and strongest verification wins the recommendation.
This shift represents a historic opportunity for brands with genuinely differentiated products but limited marketing budgets. It also poses a threat to brands that have relied on unsupported marketing claims.
"This is a one-way street," Pavey emphasised. "It's only going to become more and more important."
The question isn't whether to optimise for AI commerce. It's whether you'll help shape how your category is understood - or adapt to norms set by competitors who moved first.
Book a free 30-minute gap analysis to discover:
The future of shopping is already here -and it's powered by facts, not just marketing. The brands that embrace this reality now will define what winning looks like in AI commerce.
Oliver Farmiloe is the Growth Marketing Manager at Provenance. He works closely with our Impact team to create impactful, data-driven campaigns that help us achieve our mission of empowering consumers to make conscious, sustainable decisions online.
Provenance powers sustainability claims you can trust. The global leader in sustainability marketing technology, Provenance helps brands and retailers share credible, compelling and fact-checked social and environmental impact information at the point of sale. Provenance’s technology is already increasing conversion rates, brand value and market share for customers including Cult Beauty, Douglas, GANNI, Napolina, Arla and Unilever
.png)
%20(1)%20(2).webp)