Prada's ChatGPT app: what it does, and what it means for every other brand

What did Prada launch?

An app that lives inside ChatGPT. Prada built it with Accenture Song, with support from DaVinci Commerce, and it answers from Prada's own brand sources rather than whatever the web says about Prada. It's a pilot in three markets, the UK, the US and Italy, and Prada calls it a first for a luxury fashion house.

The quotes are what you'd expect. Laura Campanella, digital director at Prada Group, said: “Conversational experiences represent a natural evolution of traditional e-commerce.” Alessandro Diana of Accenture Song said these new touchpoints are “creating a new front door for product discovery.”

Credit where it's due: answering from your own sources is exactly right. An assistant guessing at you from third party pages is what every brand should worry about.

What can a shopper do in it?

The app's listing in ChatGPT, published by Prada S.p.A., describes a guided conversation. A shopper can browse bags, shoes, ready to wear, accessories, fragrances and beauty, see product details, pricing and purchase options, read about the ideas behind collections and shows, and find a store.

The example prompts are plain: “Show me Prada handbags.” “Where can I buy this Prada bag?” That's the whole job. Help someone choose, then show them where to buy.

The listing describes no checkout inside ChatGPT, so Prada makes the sale. One line deserves attention: once the app is connected, ChatGPT may share relevant chats and memories with it, and Prada's own terms and privacy policy govern what happens next. Planning your own app? Read that twice. The conversation arrives with context, and then it's yours to look after.

Why now?

Prada isn't first into ChatGPT, only first among luxury fashion houses by its own account. OpenAI opened apps inside ChatGPT on 6 October 2025. Louis Vuitton showed its ChatGPT app at VivaTech in June 2026, for finding products only, with purchases kept on the house's own channels. L'Oréal announced Maybelline virtual try on in ChatGPT at the same event.

Shopping inside ChatGPT has changed shape too. In March 2026 OpenAI ended Instant Checkout as a standalone product. “Instant Checkout is transitioning to apps,” a spokesperson told CNBC. Walmart had said purchases inside ChatGPT converted at a third of the rate of shoppers who clicked through to its own site.

On 1 October 2026 OpenAI added virtual try on and a library of saved products to ChatGPT's shopping results. Prada's app fits that model: it does the finding and explaining, and the buying happens where Prada already sells.

Is a single brand app just a website in a chat window?

Mostly, yes. Our read: a branded app in a chat window is a catalogue with better manners. It's fair to ask how it differs from a website. A brand's app only ever recommends that brand, and a good personal shopper would sometimes know to suggest something else.

To be fair, the manners count for something. The shopper stays in a conversation they were already having and describes what they want in their own words, instead of learning a site's menus. And the brand learns how customers describe what they are looking for. That is genuinely useful.

Here's the thing, though. Plenty of shopping questions name no brand at all: a good black bag for work, a present for someone who runs. OpenAI says people reach an app by asking for it by name or when ChatGPT suggests one. Being suggested, or named in the answer before any app is opened, depends on whether the assistant can read your products and trusts what it reads. The app is the shop window. That part is the street.

What does it mean for every other brand?

Three things, whatever your size.

1. Discovery is moving into the assistant.Prada, Louis Vuitton and the big retailers are all putting their products where the conversation happens. The shopper still buys on the brand's own site. Keep it that way.
2. Your own data is the source.Prada's app answers from Prada's own sources. An assistant that can read your product data directly doesn't have to guess from what others publish about you. At most brands that data is a mess nobody owns, which makes it the first job.
3. A custom app is the expensive way in.A global consultancy, a commerce partner, a custom build and three markets is a serious project. Most brands need the outcome, products an assistant can read and recommend, without the build.

What can a brand without Prada's budget do?

This is where we come in. Storefront does something different from Prada's app. It gives a brand its own agent address (MCP) and UCP profile, built from the product feed it already sends to Google Shopping, live in minutes. Any agent that connects can search the range and get live price, stock, sizes and images, and every result links to the brand's own product pages.

In ChatGPT, your agent address works today as a custom app, and ChatGPT shopping follows as OpenAI approves your feed. Gemini and Meta AI (Muse) shopping follow as each programme opens. The shopper buys on your own checkout, through your own payment provider, and we take no cut of your sales. It's self serve: paste the feed, start the free trial, then pay one flat monthly price.

What it won't do is design a branded conversation the way Prada's team and its partners did, or put you in ChatGPT's app directory. What it does is make your whole range readable to every agent that shops, from data you already have. It's built for brands that aren't on Shopify; on Shopify, Shopify's own agent tools cover your store.

Where do you stand today?

Before you build anything, find out what the assistants already say about you. The free score shows how the AI engines see your brand and whether shopping agents can read your products, with the fixes ranked. If you'd rather talk it through, book a 20 minute intro call.