Estimated Reading Time: 3 minutes
Key Takeaways
- AI can reduce customer uncertainty in skincare e-commerce by turning a selfie into a more personalized skin profile and routine.
- The LABRAINS approach uses small, specialized AI models instead of large mainstream LLMs or heavy data center infrastructure.
- Practical AI value comes from connecting the model to product data, customer experience, privacy, and the buying journey.
- AI should support skincare expertise, not replace it, making expert guidance more scalable and accessible.
- Tiny local AI models can lower costs, improve privacy, support data sovereignty, and offer a more environmentally friendly path for business AI adoption.
On August 22, I had the pleasure of speaking at AI Connect Latvia at RTU (Riga Technical University) in Riga, with a talk titled From Selfie to Skincare: How LABRAINS Is Doing It With AI.
It was a great event, not only because of the topic, but because of the people I met: smart speakers, curious business leaders, technical minds, organizers who kept everything moving, volunteers who made the day feel smooth, and an audience genuinely interested in where AI can create real value beyond the hype.
My talk focused on a simple business problem: skincare e-commerce gives customers endless choice, but often not enough confidence.
Anyone who has browsed an online skincare catalog knows the feeling. Cleansers, serums, creams, treatments, all promising different benefits. Filters like “dry,” “oily,” or “sensitive” help, but they are still broad. Skin is personal. Customers do not just want more products; they want guidance they can trust.
That is where the LABRAINS concept begins.
The customer experience is deliberately simple: upload face images, receive a skin profile, and get recommended day and night routines. Behind that simple flow is a more structured system: image quality checks, face detection, visible skin attribute analysis, skincare logic, product data, variant IDs, and the buying journey.
One of the main messages I wanted to share with business executives is this: AI is not the whole product. The value comes from connecting AI to a real customer need and a real business process.
In this case, AI is not just automation. It is personalization. It helps move the customer from uncertainty to a routine they can understand and act on. That creates value in three ways: a better customer experience, stronger trust in the recommendation, and a clearer path from guidance to conversion.
But building this kind of MVP also teaches humility. Applied AI lives in the details. Lighting matters. Pose matters. Image quality matters. Human nuance matters. Skincare evaluation is not a clean yes-or-no classification problem, and the AI output must be aligned carefully with expert reasoning.
Another important point from the talk was that AI does not replace skincare expertise. It can make that expertise more scalable, more accessible, and more personal. The better business question is not “Can AI replace the expert?” but “Where can AI help expertise reach more customers at the moment they need it most?”
I also shared something I believe is increasingly important: not every AI solution needs a massive data center or a mainstream large language model. The LABRAINS approach uses small, specialized models that can run locally or on modest infrastructure.
The total AI model size is around 300 MB. The MVP was created in one month.
That means lower running costs, stronger privacy, better data sovereignty, and a more environmentally friendly approach. For a focused business problem, tiny specialized models can often be enough.
That is an encouraging message for companies exploring AI. You do not always need a giant budget or giant infrastructure to start. You need a clear use case, good domain knowledge, practical integration, and a willingness to test, learn, and improve.
Thank you again to everyone at AI Connect Latvia: the speakers, the audience, the organizers, and the volunteers. Events like this matter because they turn AI from an abstract conversation into something practical, local, and human.
And that is exactly where I believe the real opportunity begins.
http://massimobensi.com/
Frequently Asked Questions (FAQ)
A: The talk explained how LABRAINS uses AI to turn customer selfies into personalized skincare profiles and product routines, helping online shoppers choose skincare with more confidence.
A: The talk was presented at AI Connect Latvia on August 22 at RTU in Riga, Latvia.
A: AI helps skincare e-commerce by analyzing visible skin attributes, reducing customer uncertainty, improving product discovery, and guiding shoppers toward more relevant skincare routines.
A: No. The LABRAINS approach uses small, specialized AI models rather than mainstream large language models like GPT, Claude, or Gemini for the skincare analysis.
A: Local AI can improve privacy, reduce dependency on large cloud infrastructure, lower running costs, and support stronger data sovereignty for customer skin analysis.
A: No. The talk emphasized that AI should support skincare expertise, not replace it. AI can make expert-like guidance more scalable, accessible, and personalized.
A: AI-powered recommendations can improve customer trust, reduce browsing fatigue, increase conversion rates, and create a smoother path from product discovery to purchase.
A: Tiny AI models can be cheaper, faster, more private, and more environmentally friendly than large-scale AI systems, especially when solving focused business problems.
A: Key challenges include image quality, lighting, face angle, dataset quality, model fine-tuning, product catalog integration, and ensuring customers can trust the results.
A: Executives can learn that practical AI success depends less on hype and more on clear use cases, domain knowledge, customer trust, operational integration, and measurable business value.
