Putting generative AI to one side, in what ways, specific to your area of expertise in the beauty industry, are you seeing AI being used? Can you share some AI success stories?

If there’s one thing we’ve learned working with beauty brands, it’s this: more choice doesn’t necessarily drive growth. Better decisions do.

Across categories, the same pattern keeps showing up. Consumers aren’t struggling to find products - they’re struggling to choose between them. And that hesitation has a real cost: abandoned carts, returns, and low confidence in online purchases.

Brands should be guiding consumers to the right choice with clarity and confidence. That’s where beauty tech actually proves its value. Not in how advanced it looks, but in how effectively it helps people decide.

AR Try-On Apps

AR try-on solved one problem very well: hesitation. But the conversation shouldn’t stop there.

What’s often missed is that most AR experiences still behave like entertainment layers, not decision tools. They look good, they drive engagement, but they don’t necessarily help users make a confident choice.

That gap comes down to realism.

If the finish is off, if undertones shift, if textures like shimmer or matte don’t behave as they would in real life, the experience breaks. Consumers might still play with it, but they won’t rely on it. And if they don’t rely on it, it doesn’t influence the purchase in a meaningful way.

This is where the real divide in AR sits today: between visualization and simulation.

Visualization shows a product on a face. Simulation reflects how that product actually behaves on your face, with your skin tone, lighting, and features. That’s much harder to get right, but it’s also where the commercial value is.

At Arbelle, this is exactly the problem we’ve focused on solving. Not just placing color on a face, but accurately rendering finishes, multi-shade products, and skin tone alignment without “beautifying” or distorting the user’s appearance. Because once you start smoothing, filtering, or subtly altering the face, you’re no longer helping the user decide - you’re just making the result look better than reality.

And consumers are increasingly sensitive to that.

There’s also a second layer that’s often overlooked: measurement.

Many brands still evaluate AR based on engagement metrics - time spent, interactions, shares. But those are weak indicators of business impact. The more relevant question is: does it reduce returns, increase conversion, and improve repeat purchases?

When AR is connected to analytics, it becomes much more than a front-end experience. It becomes a source of insight. You can see which shades are tried most, where users drop off, what drives clicks, and how different segments behave.

That’s where the real shift happens - from AR as a feature to AR as a decision engine.

And this is also how it bridges digital and physical retail more effectively than people expect. Consumers don’t just “try for fun” anymore. They use these tools to pre-select, narrow down options, and arrive either online or in-store with intent.

In that sense, AR isn’t replacing physical retail. It’s quietly reshaping it by moving the decision-making moment earlier in the journey.

Many shoppers now arrive in-store already knowing what they want to try, or they skip the store entirely because they feel confident enough to purchase online.

The bottom line: AR try-on works when it earns trust. And trust comes from accuracy, not aesthetics.

Personalized Product Recommendations - AI-driven suggestions for cosmetics

Personalization in beauty is often oversold and underdelivered.

We’ve seen firsthand that there’s a big difference between asking a few quiz questions and actually guiding someone to the right product. Consumers can tell the difference - and they trust it accordingly.

The more advanced systems combine multiple signals, but more data doesn’t automatically mean better recommendations. What matters is how well that data connects to real decisions.

From Arbelle’s analytics, for example, this becomes very clear. You can see which shades and products are tried most, where users drop off, what drives a click, and how different skin tone groups interact with specific products. That’s where personalization becomes actionable - not just suggesting products, but understanding what actually leads to a confident choice.

We see this most clearly with foundation shade matching. When it’s accurate, it reduces returns, increases conversion, and builds repeat purchases.

Personalization works when it removes doubt. If it adds complexity, it’s doing the opposite of what it should.

Beauty-Tech Integration

Most beauty tech still exists as standalone features. That’s where the friction comes from.

But the real value actually comes from connecting key touchpoints, such as virtual try-on, shade matching, and product recommendations into one continuous experience. Not separate steps, but a guided flow that helps users move from discovery to decision without guesswork.

This is also where analytics becomes critical. When these elements are connected, you can see what actually drives decisions - what leads from try-on to selection, where users hesitate, and what converts.

Integration isn’t about adding more. It’s about removing gaps between steps.

Because from a consumer’s point of view, there are no “features.” There’s just one question: does this help me choose?