Every year at the National Retail Federation’s Big Show, the show floor fills with hundreds of booths from tech companies selling the next big thing in retail innovation. This year’s show, which ended in mid-January, was no different — and the next big thing on the show floor was AI agents.
That was confirmed last week when OpenAI announced the imminent release of Operator, its take on the AI agent formula. But what is an AI agent and how might it affect the fashion and beauty industries? Glossy spoke with experts in AI tech and business to get the info.
What is an AI agent?
An AI agent, essentially, is a tool that uses a large language model, the same technology that powers AI chat programs like ChatGPT or Google’s Gemini, to browse and perform actions on the internet. The term “agentic” is used to refer to AI programs with the ability to take actions, while non-agentic refers to simple chatbots or AI programs that can do things like respond to prompts but cannot actually act on those prompts externally.
“An AI agent is software that can perform tasks, make decisions, and interact with users or systems without constant human supervision,” said Nathan Brunner, CEO of Boterview, a company that makes an AI-powered interview preparation program. “Think of it as a digital assistant powered by technologies like computer science and natural language processing. It’s like ChatGPT, but tailored for specific purposes.”
For example, while users can ask ChatGPT for recommendations on good restaurants in their area or a fun place for a camping trip, Operator can venture out into the internet and make a dinner reservation or buy camping supplies.
How do they work?
AI agents can “see” the web through taking screenshots of websites and analyzing what elements it detects on screen. Then, it can perform any action that a user could with a mouse and keyboard. Most will pause and ask for further directions if stuck or require user input for sensitive actions like entering credit card information or passwords and making purchases.
What AI agents are out there?
The aforementioned Operator made headlines for its open research preview on January 23, before launching in the U.S. soon after at OpenAI’s $200 per month price tag.
But OpenAI isn’t the only player in the AI agent race. ByteDance, the owner of TikTok among other companies, has a model called UI-TARS that it made open source the day before Operator was announced. UI-TARS works on a similar model, recognizing elements on the screen and taking actions based on its initial prompt.
While Operator and UI-TARS are general purpose agents, there are dozens of niche agents for use in backends for things like tagging and organizing product information.
“Google Dialogflow is a big name; it’s like the go-to for chatbots,” said Volodymyr Murzak, solution architect and tech lead at software development company Syndicode. “Microsoft’s Bot Framework is great if you’re in the enterprise zone. For something more customizable, there’s Rasa. It’s open-source and lets you tweak things just the way you want. Then, there are tools like Vue.ai that are built explicitly for the fashion biz. They handle stuff like styling advice and tagging products.”
How can brands make use of AI agents?
Customer use of general purpose agents like Operator won’t require any change on a brand’s part. The agent would theoretically interact with the website in exactly the same way a user would.
But there a number of ways brands could make use of agents themselves. The aforementioned Vue.ai has lots of potential applications from tagging product information, creating virtual fitting rooms, developing thematic product curations and facilitating dynamic pricing. Vue is already in use by companies like ThredUp, Nike and Levi’s.
Other fashion brands are using AI agents for counterfeit verification. Lacoste has made use of a tech called Vrai AI, made by the French company Cypheme, that can detect counterfeits.
“The next step would be to create an AI agent that scours the internet to combat counterfeits by scanning product images to detect fakes,” Brunner said. “This would be ideal for protecting the brand’s reputation and combating unlicensed resellers.”
Christian Hed, CMO of the European communications tech company Dstny, told Glossy that perhaps the best use of AI agents for fashion brands would be in trend analysis.
“The fashion world is always about staying on the cutting edge — knowing what’s happening, what people are posting, what trends are emerging and even the latest gossip — so brands can iterate, stay relevant or even get ahead of the game,” Hed said. “AI agents are absolutely brilliant tools for trend analysis. For example, you can use them to monitor all the top magazines in the fashion world, collect top stories of the day into an Excel spreadsheet and pull out themes from the titles. They can also monitor Reddit accounts, forums or social media, staying on top of different trends.”
Should you use an AI agent?
A common theme in the age of AI is an immense hype bubble that inevitably bursts when the reality is less than the idealized sales pitch. It’s been seen with Apple’s AI venture, Apple Intelligence, which failed to excite customers and generated controversy for erroneously summarizing news headlines. AI-powered hardware like Humane’s AI pin flopped, and AI chatbots and search results infamously and regularly give out false information.
Goldman Sachs further threw cold water on the hype last summer with a report on the lack of visible returns on the AI gold rush.
Melanie DiSalvo, founder of apparel manufacturer Virtue + Vice who has worked with major clients like Rag & Bone, Walmart, Ralph Lauren and Target, told Glossy she would suggest avoiding AI agents.
“While AI is great for scraping data like emails, it’s not great for closing sales,” she said. “When it comes to wholesale orders, 90% of the deal is the relationship. There is no shortage of new brands for boutiques and department stores to choose from. Where brands stand apart from all the others is the personal experience and relationship they can create.”
Other users of AI agents have touted their success. ThredUp founder and CEO Chris Homer called Vue.ai a “very powerful addition to our platform.” But even among proponents of the technology, there’s a sense that businesses should be cautious about being wowed by shiny objects.
“The trick with AI agents is to use them to solve real headaches, not just because it sounds fancy,” Murzak said. “They’re excellent at speeding up workflows and cutting out repetitive stuff, but you don’t want to overdo it and create more work for your team. Balance is everything.”