Chrome Photos

Eduardo Jacqui Alps, 64, built Chrome Photos to solve a massive inefficiency in e-commerce: sellers struggle to know what product image elements actually drive sales. He saw the opportunity to combine data science with a global network of designers to create images that definitively lift conversion rates. The company was "launched a few years ago" but the core insight—that AI-driven image creation at scale could replace fragmented freelance platforms like Fiverr and Upwork—came from observing that e-commerce is fundamentally broken when it comes to visual assets.

The team built a two-layer system: software and humans. Everything before image creation is AI and machine learning—analyzing what elements (colors, composition, models, backgrounds) move the needle for specific products. Then, when a customer sends in a product, generative AI pre-assembles the image 60-70% of the way, and human editors finish it in Photoshop. They structured this like Uber: 200 editors and designers worldwide receive job requests with clear pay ($5-10 per image) and specifications. By using Adobe XD for design lockups and flowing into the editing team, they created a repeatable, scalable assembly line.

Chrome Photos landed marquee enterprise customers—Amazon (top-ranked Amazon SPN seller), Walmart (helping national sellers), and Subidis (for online auctions)—likely through direct enterprise sales. They've now processed over 1 million images for over 100,000 customers total. What made them attractive: they could prove with data that their images lifted sales, something ChatGPT or generic AI image generators cannot do. Customers pay per image ($25-50) or via subscription credits (similar to Audible), with most using the service for lifestyle photos and infographics.

The biggest insight: generic AI images are not good enough. ChatGPT succeeds with text because everyone can edit text; most people cannot edit images, and they don't know what product elements drive sales. Chrome Photos' moat is data—they have 8.4 million data points on what sells—plus the ability to take generic AI outputs and perfect them with design expertise and contextual understanding. As generative AI improves, they're automating more of the human work, with images becoming 70%, then 80%, then 90% complete before human touch, which will eventually push their 55-60% margins toward SaaS-like 85%+. The risk they acknowledged: competitors with deeper funding or ChatGPT's own vision models could replace them entirely—but today, that's not happening.

In March (the interview's "last full month"), they processed 15,000 images across 500 paying customers, generating over $200,000 in revenue, up from $100,000 a year prior. They closed a $600,000 seed round from Evolution VC (lead: Greg Smith) at somewhere between $10-20M post-money valuation—a very capital-efficient raise where Eduardo kept "much less" than 10% equity dilution. The capital powers their Walmart partnership (Walmart is flagging 50M products that need updates and directing sellers to Chrome Photos, expecting 20-30% conversion). They're targeting $400-600k MRR by December and planning a Series A in Q3. With just six full-time employees and a gig network handling execution, they're positioned to capture what they estimate is a $60B+ market—currently fragmented across Fiverr, Upwork, and scattered studios. Eduardo declined a hypothetical $10M offer, saying the real opportunity is exits of $100M+.