Look, a Google-backed client tech firm, launched Look AI on Thursday. This new app permits customers to strive on AI-curated outfits utilizing only a selfie.
The app recommends vogue kinds that go well with every consumer’s look, physique kind, and preferences. Customers can discover over 400 world manufacturers and store straight from the app.
“The web is being rewritten from curated feeds to AI-generated realities…This is without doubt one of the most formidable technological efforts in our journey and a daring step towards shaping the way forward for AI-driven client experiences,” Naveen Tewari, founder & CEO of InMobi and Look, added.
Not like conventional e-commerce apps, customers don’t have to scroll by means of limitless lists of merchandise. As an alternative, Look AI generates seems which might be styled particularly for them and exhibits how they’d seem in actual life utilizing AI fashions.
The corporate claims that 1.5 million customers within the US have tried the app, and almost half changed into repeat clients. To this point, over 40 million model requests have been made, and 40% of customers faucet by means of to buy after seeing their seems, Look stated.
The app provides folks management over their knowledge whereas providing an expertise that feels private and interesting.
Why not use in-house LLM’s?
In an interview with AIM, Mohit Saxena, co-founder and CTO of InMobi and Look, said that with foundational fashions like GPT, Claude, and Gemini developed over a decade of analysis and billions in funding, there’s little incentive to “reinvent the wheel.”
Thus, the corporate is leveraging these highly effective fashions—typically open-source like DeepSeek and different industrial fashions like Gemini Professional and fine-tuning them for particular, localised use circumstances.
Saxena talked about that the actual mental property (IP) lies in “knowledge preparation, contextual layering, and post-processing.” Consequently, this multi-model orchestration is prone to stay the corporate’s strategy going ahead. No single mannequin immediately can grasp each step within the chain, Saxena added.
Nevertheless, there are limitations. Most foundational fashions are predominantly educated on Western knowledge, which leads to a lack of information of Indian and broader Asian cultural nuances.
This underscores the necessity for contextual AI tailor-made to India. “Localisation is the important thing. We’d like AI constructed for India and educated in our languages, dialects, and cultural contexts. That’s the future I consider in,” Saxena added.
Bengaluru Staff Behind the Scene
Look AI’s proprietary tech stack was constructed solely by the engineering staff in Bengaluru.
Operating superior AI fashions sometimes requires costly GPUs. Within the early levels, producing simply 20 photos may price over $100, Saxena revealed. Nevertheless, by means of “important engineering effort,” the staff optimised the method, dramatically decreasing prices whereas sustaining excessive output high quality.
Beforehand, making a personalised model profile required round 20 photos per consumer. A big innovation modified that. The system now works with only a single selfie.
Whereas India stays an necessary market, its core enterprise spans throughout the US, Europe, China, and different areas.
“That is the primary platform to be accessible concurrently on the lock display, ambient screens, and as a standalone app, providing a seamless and distinctive consumer expertise throughout units worldwide,” Saxena said.
Look AI has partnered with over 400 vogue manufacturers, together with Ajio, Zara, and ASOS, with plans to increase additional. Whereas the platform is launching specializing in quick vogue, it’ll quickly department into different classes corresponding to sneakers, eyewear, cosmetics, make-up, way of life merchandise, and residential decor.
The publish Look Launches AI Purchasing App to Improve Client Expertise appeared first on Analytics India Journal.
