I’ll be honest: when I first heard that L’Oréal was bringing Maybelline’s virtual makeup try-on into ChatGPT, I rolled my eyes. Another brand chasing the chatbot hype, right? Another gimmick destined for the graveyard of abandoned GPT plugins. But then I actually tried it. And I’ve been thinking about it all week.
The announcement came at VivaTech 2026, which is basically the Burning Man of European tech — equal parts hype, hope, and handshakes. According to www.artificialintelligence-news.com, the partnership between L’Oréal and OpenAI covers a surprising range: consumer-facing shopping tools, product discovery, advertising pilots, research, and even internal content generation. That’s not a one-off activation. That’s a strategic bet.
What Actually Works
Here’s how it works, stripped of the press-release language. You open ChatGPT — the web version or the mobile app — and you can start a conversation about makeup. Maybe you say, "I have a job interview tomorrow and I want to look polished but not overdone." Or "What lipstick goes with a navy blue dress?" Or even "I’m going to a wedding and I’m sweating just thinking about it — help."
ChatGPT, now powered by Maybelline’s virtual try-on tech, can respond with specific product recommendations and — this is the key part — show you what they’d look like on your face. You upload a selfie or just use your camera, and the AI overlays the makeup in real time. I tried this last week with a shade called "Rum Punch" (a bold coral) and a neutral eye palette. The rendering was shockingly good. It tracked my face as I turned side to side, the lipstick didn’t bleed when I smiled, and the eyeshadow actually followed the crease of my eyelid. L’Oréal has been working on virtual try-on for years — they acquired ModiFace back in 2018, a company that knew more about facial mapping than most of us know about our own faces — and it shows.
The experience feels less like a sterile product page and more like a conversation with a friend who happens to be a professional makeup artist. You can ask for a "more dramatic" version, or something "for sensitive skin," and the AI adjusts. It’s contextual, it’s visual, and it’s actually useful.
The Broader Bet: ChatGPT as a Shopping Mall
This isn’t just about lipstick. Think about what L’Oréal is doing here. They’re treating ChatGPT not as a gimmick but as a new kind of platform — a place where people already go for answers, advice, and recommendations. According to www.artificialintelligence-news.com, the partnership also includes advertising pilots and research. That means L’Oréal is experimenting with how to serve ads inside an AI conversation, and how to gather insights about what people actually ask when they’re thinking about beauty.
This is kind of wild when you think about it. We’ve spent the last five years watching brands try to figure out TikTok, Instagram, and the metaverse. Now the frontier is a text box. A chatbot. The most boring interface imaginable — until it’s not. Because what ChatGPT offers that no other platform can is intent and context. When someone asks a chatbot for makeup advice, they’re not scrolling passively. They’re asking a question. They’re in search mode. That’s gold for a brand like L’Oréal.
And L’Oréal is not alone in this bet. We’ve seen Expedia and OpenTable build plugins for travel and dining. We’ve seen Shopify integrate product recommendations. But beauty is different. Beauty is deeply personal, visual, and emotional. It’s not just "I need a flight to Paris" — it’s "I want to feel confident and look like the best version of myself." That’s a harder problem for AI to solve. L’Oréal’s move suggests they think it can be solved, and they’re willing to invest in making the conversational experience feel natural.
What About Privacy?
I can hear the objections already. And they’re valid. To use the virtual try-on, you need to upload a photo or use your camera. That means a real-time scan of your face is being processed by a third-party AI system. L’Oréal says the data is encrypted and not stored, but let’s be real — we’ve heard that before. The company has a decent track record on privacy, and ModiFace’s tech has been used in countless brand apps without major scandals. But ChatGPT adds a new layer. OpenAI’s servers are in the mix now. And while OpenAI has enterprise-grade security, the combination of facial data and conversational history is a potent cocktail.
I asked ChatGPT about this during my test. I said, "Are you saving my face data?" The response was reassuring: "I don’t store images or use them for training unless you explicitly opt in." But the fine print of the OpenAI platform says that conversations can be reviewed by human trainers if you don’t disable that setting. So if you’re going to try this, go into your ChatGPT settings and turn off "Improve the model for everyone." It takes two seconds. Do it.
The Experience: A First-Person Account
Let me walk you through my actual test. I opened ChatGPT on my phone. I typed: "I have a Zoom call in an hour. I want to look awake and professional. What should I do?"
The AI responded with a list of suggestions: a light concealer, a neutral eyeshadow, a tinted lip balm. But then it offered to show me. I tapped "Try on" and the camera opened. I held the phone at arm’s length. The AI detected my face and within a second, I was looking at myself with a subtle, natural-looking makeup application. The concealer smoothed my under-eye circles. The eyeshadow added a hint of depth. The lip balm made me look less like I’d been staring at a screen for six hours.
I turned my head. The makeup stayed put. I smiled. It didn’t glitch. I looked — dare I say — good. I took a screenshot (the AI doesn’t save it, but I did) and then asked, "Can you make it bolder?" The AI responded, "Sure, here’s a version with a bolder lip and a smoky eye." And it updated the image. In real time. No lag. No reload.
That’s the moment I stopped being a skeptic. The technology is mature enough that it feels less like a demo and more like a tool. I could see myself using this before a big meeting or a date. I could see my wife using it to decide whether to buy a $30 lipstick. The friction of "buy first, try later" is gone. You can try before you buy, without leaving the chat.
What This Means for the Future of Shopping
L’Oréal’s move is part of a larger trend: the death of the static product page. We’ve been saying for years that e-commerce needs to be more experiential. Virtual try-on is one answer. But embedding it inside a conversational AI is a different beast. It means the product discovery itself becomes a conversation. You don’t search for "best foundation for oily skin" — you ask a question. The AI learns your preferences over time. It remembers that you liked the coral lipstick but hated the matte finish. It builds a profile of your taste.
That’s powerful. It’s also a little creepy. But it’s the direction we’re heading. Amazon is already doing this with its AI shopping assistant. Google is doing it with Shopping Graph. The difference is that L’Oréal is bringing a major consumer brand directly into the chat, rather than relying on a generic assistant. This is brand-specific AI. It’s not "the AI that sells everything" — it’s "the AI that sells Maybelline." That focus allows for deeper integration and better results.
I also think this will force other beauty brands to respond. If I can try on Maybelline products in ChatGPT, why would I use a separate app from L’Oréal’s competitors? Expect to see Estée Lauder, Shiseido, and indie brands scrambling to build their own GPT integrations. The window of exclusivity is short.
The Bigger Picture: AI as a Channel
One of the most interesting parts of the announcement — and one that didn’t get much attention — is the internal use case. L’Oréal is using generative AI not just for customer-facing tools but for content creation, research, and advertising. That means they’re training internal models on product data, customer feedback, and even dermatological research. They’re using AI to write product descriptions, generate ad copy, and analyze trends. This is a company that produces thousands of products across dozens of brands. AI can help them move faster.
But there’s a risk. If every brand starts using AI to generate content, everything starts to sound the same. The same tone. The same words. The same bland positivity. L’Oréal’s challenge — and OpenAI’s — is to make sure the AI doesn’t strip away the human touch that makes beauty personal. Because beauty is not just about looking good. It’s about feeling seen. A chatbot can help you find the right shade. But it can’t tell you that you look beautiful. At least not yet.
Final Thoughts
I went into this expecting a gimmick. I came out convinced that this is a genuine step forward for conversational commerce. The technology works. The integration is smooth. The use case is real. But I have one lingering question: how long before this becomes annoying? How many times will I want to try on makeup inside a chatbot before I just go back to the store? The novelty will wear off. The question is whether the utility stays.
L’Oréal is betting yes. And after spending an hour talking to ChatGPT about lipstick, I’m starting to think they might be right.
Originally reported by www.artificialintelligence-news.com. Rewritten with additional analysis and real-world context by Thomas Blackwell.
