FT Live 2026: Future of Retail review: unlocking the next phase of agentic commerce for online stores
“Your product page is your new home page,” when it comes to discovery in the agentic AI era, said Philip Hall, MD, Europe at Rithum, during FT Live 2026: Future of Retail in London last week. Fellow conference session panellists from L’Oréal and Wolf & Badger, a marketplace that champions independent fashion and homeware, emphasised how important good data tagging, quality, and hygiene is amid a digital revolution where AI agents can buy on behalf of consumers.
According to Matt Berzinski, EMEA Field CTO at Ping Identity, who was also on the panel entitled ‘The new online storefront - unlocking the next phase of agentic commerce’ on 6th October: “AI agents are inherently lazy, so if you’re too hard to find they will just move on.” Having good data mastery is therefore more important than it has ever been.
“It is also up to us to identify agents and control them,” he added, as the conference session panel moved on to discuss the centrality of good governance and know your agent (KYA) compliance. This could become as vital as customer (KYC) compliance requirements in the future. “Trust is crucial in this coming era where the ability to know who is a human, who is an agent, and to have appropriate verification is key.”
GEO
Aside from security internally and with ecosystem partners, data mastery is also vital for the Generative Engine Optimisation (GEO) future that will displace traditional SEO in search in the future. The skill of structuring and tailoring online content so that AI large language models (LLMs), such as ChatGPT and Gemini, can find and recommend your brand will be crucial in driving future sales.
As Othman Bennis, Chief Digital and Marketing Officer, L’Oréal Europe, said: “L’Oréal is moving to having 18-20 data fields to hundreds in the future in terms of our product attributes. This is necessary for the agentic AI future.” He also shared that his firm is working with OpenAI now that they’ve opened an ad function on ChatGPT.
Early days
George Graham, CEO at Wolf & Badger, noted that his firm was the first in Europe to take an agentic AI initiated order. “We are already transacting,” he said, while stressing it is very early days in the migration to agentic commerce. “We use Stripe’s agentic commerce suite of tools and work with Meta to enable our offering.”
Pricing pressure
In regard to the potential pricing pressure from AI agents searching out better deals for their humans, using economies-of-scale negotiating tactics depending on their number of adherents, Rithum’s Hall thought this likely. “LLMs will tend to direct you to the cheapest price. However, I personally don’t just shop on price, so there is still room for retailers to differentiate themselves.”
According to Wolf & Badger’s Graham fundamentally they’re still a retailer that needs to curate their offering to achieve the best products on their platform, the best service, loyalty and retention, at the best price they can get. This means delivering: Good CX; Products, price and service; Fulfilment. “These are always at the core of good retailing,” he said. “The fundamentals don’t change.”
L’Oréal Europe’s Bennis agreed, although he did stress his company is not complacent about the coming revolution and is already working with Alibaba and Amazon among others to protect discoverability. “Ultimately though, it is the one with the best CX that will win this race.”
There is no doubting that the race is underway. “Don’t leave too many data gaps and you’ll be OK,” he reassured attendees, while stressing that if you do then the agent might get it wrong when classifying your brand. “Don’t give them the chance. Don’t put your brand out there with a terrible dataset.”
It appears that data really is the ‘new oil’ in the coming agentic AI era where it will grease the modern economy in the way petrol did the 20th century. Retailers need to be ready for the coming revolution.
Q&A
In response to questioning from RTIH, the participants said they weren’t worried about the possibility of AI agents sifting through data that had already been sifted by other agents.
The virgin internet, where human checked and verified information can be mined and interrogated, is fast becoming a thing of the past. But this wasn’t a worry to the panellists, or at least it’s outside of their control and parameters. Verification against inaccuracies or fakes is also a potential countermeasure.
“Even if you’re training on synthetic data that has been artificially created, retailers will still have their own data control mechanisms and procedures,” said Wolf & Badger’s Graham. Rithum’s Hall, meanwhile, advised retailers in the room to just concentrate on what they can control and to get that right.
