Lumesa, a New York startup building artificial intelligence tools that generate product images for online retailers, secured $6 million in seed funding led by Work-Bench, the company announced February 5. The round included January Ventures, Red Bike Capital, Outlander VC, AI Furnace and several strategic angels, bringing total capital raised to $7.5 million.
The company, which rebranded from Flock AI during 2026, trains proprietary models on individual brand aesthetics to produce product shots, AI-generated models and virtual try-on features. Its "Brand DNA system" analyzes more than 200 visual attributes specific to each retailer, according to the company's website. Current integrations include Shopify, Shopify Plus and BigCommerce.
One enterprise customer generated between 5,000 and 6,000 images in a single month and recorded roughly a 3 percentage point increase in conversion rates, according to company and investor-reported data. Work-Bench specializes in enterprise software investments.
Lumesa's case studies page cites a multi-brand apparel holding company valued above $1 billion that doubled production volume over six months without physical samples or traditional photo shoots. A global luxury accessories conglomerate saw a 3 percent conversion lift, the company reported. Those case studies remain anonymized and rely on company-provided data.
The startup claims its platform delivers tenfold faster production, cuts content costs by 80 percent and lifts conversions between 20 and 30 percent—internal estimates reported by the company rather than independently verified figures.
Retail Veterans Turn to Generative AI
Co-founder and CEO Manvitha Mallela previously worked in merchandising and creative operations at Bloomingdale's, with additional stints at Jet.com and Walmart, Sourcing Journal reported in June 2024. Co-founder and CTO Malavika Reddy built AI and computer vision systems at Microsoft before leading software engineering at Brightcom, according to the company's blog.
"Creative teams need leverage," Mallela said in the press release. The platform uses reinforcement learning to train custom models while keeping "creative judgment with humans," she added.
Lumesa lists between 11 and 50 employees on its website. LinkedIn showed 52 staff members in an earlier September snapshot, though exact timing remains unclear.
The company said in a May blog post it would channel the fresh capital toward building "scalable sales and marketing infrastructure." Management declined to disclose valuation.
Competitive Landscape and Fashion Tech Credentials
Work-Bench general partner Jonathan Lehr told Vogue Business in January that "Flock is becoming the content infrastructure for modern commerce," using the startup's former name. His firm has backed multiple enterprise software companies navigating the transition to AI-native workflows.
Lumesa operates in a market where consolidation has started to reshape the competitive field. Browzwear, a 3D apparel design platform, acquired Lalaland, another AI virtual models startup, though the timing and terms of that deal remain publicly vague.
The company earned a spot in the 2024 New York Fashion Tech Lab cohort, which focuses on women-led B2B retail technology ventures. That program has become a recognized credential in the retail tech ecosystem.
Lumesa initially raised $1.5 million in a pre-seed round led by Outlander VC in June 2024, Sourcing Journal reported. Backers included Red Swan Ventures, F4 Fund, Hustle Fund, Dorm Room Fund and The MBA Fund. Outlander returned for the seed extension.
Yahoo Tech mentioned the startup in September coverage of Anne Klein's virtual try-on feature, highlighting Lumesa's size-inclusive approach to AI-generated models and digital twins. The company addressed Amazon's planned 2026 AI image tagging requirement in an August blog post, noting sellers must flag "contains-synthetic-performer" for fully AI-generated photorealistic people under rules tied to forthcoming New York legislation.
That regulatory backdrop suggests the technology is maturing faster than compliance frameworks can keep pace with, a dynamic familiar to anyone tracking generative AI deployment in consumer-facing sectors.
