Agentic AI is the latest tech buzzword on every fashion exec’s lips. The term dominated discussions at Cannes Lions; while at Vivatech, LVMH chief information officer Franck Le Moal announced the conglomerate is developing artificial intelligence agents in partnership with Google. At Shoptalk, Diane von Furstenberg CEO Graziano de Boni told Vogue Business that the company is considering how to use AI agents.
Agentic AI is a step beyond the conversational AI applications like ChatGPT that brands have been using for a while now. Where these large language models (LLMs) provide responses to human prompts, AI agents are software systems built on top of LLMs to independently complete tasks on behalf of humans. They can autonomously set goals, make decisions and handle open-ended, complex projects.
Their uses for fashion brands include prompt-and-response agents that act as much better customer chatbots that do away with any need for human intervention — see the launch this week of Daydream, a new fashion discovery platform that uses an AI agent to understand customer queries and recommend products. They could even buy products on customers’ behalves, with the right programming.
Founder Julie Bornstein raised $50 million last year to create the ultimate AI shopping experience. Today, it launches in beta.
“The value of agentic AI for retail is you could create efficiencies and making decisions quicker, to free up resources to be more strategic,” says Anjali Burkins, senior director of North America retail strategy at O9 Solutions, which works with brands including Prada Group, J Crew, Pandora and Ralph Lauren on backend operations like supply chain planning.
While agentic AI has potential for fashion, it’s not just yet arrived. “The foundational technologies for truly agentic AI, such as better models, tool use, longer context windows and improved reasoning, are advancing rapidly,” says Matthew Drinkwater, head of the Fashion Innovation Agency at London College of Fashion. “However, the most sophisticated, fully autonomous agents are still some way off from reliable deployment in 2025.”
Changing customer interactions
The potential of AI agents for fashion brands breaks down into what experts consider ‘invisible’ and ‘visible’ intelligence.
Invisible intelligence happens behind the scenes, requiring no interaction with customers. Last week, Le Moal said the company is developing task agents using Google’s Vertex AI platform that sales advisors at some of its 75 brands like Celine and Tiffany’s could in the future consult for fast, detailed information about the customers that visit them in-store — invisible intelligence that acts like a digital clienteling 2.0.
“It’s about weaving together data and AI that connects the digital and store experiences,” Le Moal said. “We want our customers to get even more dedicated attention without being intrusive.” These agents aren’t replacing sales advisors — instead, LVMH is experimenting with the technology to help them deliver an even more ‘human’ touch.
“We call it invisible intelligence because it’s a way to empower sales associates to provide the best service at a much larger scale,” says Raakhi Agrawal, managing director and partner at management consultancy BCG. “Currently, brands do this for their top 1 per cent of customers, or their VICs. But now, they’re thinking about how to scale that to the next 30 per cent of their highest-spending customers.”
This intelligence is being described as agentic, because the AI task agent can quickly scan through a much larger set of complex data than before, making connections and then proactive suggestions. It may suggest the sales advisor offer Client X a private appointment at their boutique on Tuesday next week, when the item they tried on last week is due to be restocked in their correct size.
Visible intelligence happens in direct customer interactions. Tech investors are backing a new wave of AI-powered personal styling agent startups that are hoping to change the way consumers shop by connecting them to brands via customisable virtual wardrobes, and taking an affiliate cut. The latest example is US-based Alta, which raised an $11 million seed round led by Menlo Ventures last week.
Alta’s app, currently in public beta mode, lets users create their own personal avatars and upload items from their real-life wardrobes to their virtual wardrobe inventories. They can create wishlists and specify their favourite brands, and ask the AI agent what they should wear for a specific occasion, while giving it feedback when it throws up a suggestion they wouldn’t wear.
“What makes this agentic is this autonomous future suggestion component,” says Alta founder Jenny Wang. “Our AI model is constantly parsing your feedback, learning from it, then adding it to what it knows about you so it can surface better recommendations for you going forward.”
Wang says she wants to explore deeper integrations with brands in the future, where exclusive capsule collections and drops could also be showcased on the app.
Internally, agentic AI could make it possible for brands to develop their chatbots so that customer queries can eventually be handled end-to-end without the need for any human intervention. Saks Fifth Avenue is one brand that says it’s one step closer with the latest iteration of its customer service chatbot Agentforce, which it launched in September last year in partnership with Salesforce.
“We’re seeing early indications of higher containment rates — autonomous resolution of customer inquiries — compared to our previous chatbot,” says Saks Global chief technology officer Mike Hite. “It can also take actions to provide more proactive and faster resolutions to supplement the enhanced contact generation seen from leveraging LLMs.”
Developers say it’s in brands’ interests to invest in personalising these systems to match with their brand DNA. “Brands are highly incentivised to get people to interact with their own agents so they can cultivate their own brand voice,” says Fayez Mohamood, CEO of AI shopping assistant Alby’s developer Bluecore.
Elsewhere, companies that are developing AI shopping assistant software to sell to retail are experimenting with customer service agents that could continue the customer conversation offline or on Whatsapp, or even complete a purchase at checkout. In January this year, Open AI released its Operator business agent — currently available to US business customers in research preview mode — which can fill out forms and order groceries, or book trips. It plans to sell this product directly to its paid consumers in the future. Although these ideas are still quite nascent, the space is moving fast.
However, in luxury fashion — where so much importance is placed on white glove service until the purchase decision — it’s easy to see why analysts say this is one of the areas that brands are the most cautious.
“If you have an agent providing ‘personalised shopping advice/suggestions’, and it goes seriously off-piste, at this higher average order value, it could really damage the brand experience,” says Tanya Goodin, founder and CEO of EthicAI.
Movement behind the scenes
Although only a handful of brands have spoken out about AI agents at customer touchpoints, analysts say that the majority of luxury brands are already investing in behind-the-scenes use cases.
“In areas like trend forecasting, the increased agentic characteristics of the AI systems mean we’re seeing more automated forecasting in natural language from the AI itself, so that buyers aren’t having to rely so much on their own gut feel for the future” says Burkins.
“In the past, these business functions have been quite siloed, but now these agentic capabilities mean the AI systems are creating dashboards and giving planning teams answers to questions much faster so they can make more reactive decisions,” Burkins says.
The ultimate goal with using AI agents for supply chain management is to move towards completely touchless planning. One of O9 Solutions’s clients has already moved to 90 per cent touchless planning, Burkins says, thanks to its use of AI agents for nearly all its forecasting.
“When you think about the learning curve of bringing in AI agents into organisations, it’s definitely a crawl, walk, run process,” she says. “And this client I mentioned is definitely in its run, maybe even rocket ship, stage compared to most other fashion brands.”
Marketing experts also say that luxury and fashion brands are experimenting with AI agents for initial creative briefings, trend forecasting and mood board generation.
Generative AI has already been able to analyse data like fashion shows, streetwear blogs, social media and retail trends to generate summary visual mood boards and style trend reports. But now, brands are using AI agents for proactive concept ideation based on a brand’s DNA.
“Brands are using custom-trained task agents that understand their brand voice, past campaigns and audience preferences to suggest initial campaign themes, story arcs, product ideas and even early-stage visual and copy prototyping,” says Trey Courtney, global chief product and partnerships officer at Mood Media.
One brand that’s experimenting with AI agents for marketing is The Estée Lauder Companies (ELC), which has partnered with Microsoft to develop its ConsumerIQ agent. The prompt and response agent is able to analyse the company’s 80 years’ worth of archives and data to quickly surface the most relevant insights for its marketing teams when they’re thinking about how to market and develop new products.
Yet when it comes to editorial content, analysts say that’s where most luxury brands still draw a line — and where humans will remain in the loop for the foreseeable.
“Brands are still going to use their trusted photographers and models and film and shoot true to the brand,” Agrawal says. “Their aesthetic is so powerful and important for telling their brand story that we’re seeing they’re far less likely to experiment there.”
Sign up to receive the Vogue Business newsletter for the latest luxury news and insights, plus exclusive membership discounts.
Comments, questions or feedback? Email us at feedback@voguebusiness.com.
Is Daydream’s AI platform the answer to fashion’s discovery problem?
Gen Z is using ChatGPT as their stylist. What does it mean for brands?
