Lumesa

Cover Block

Open sources

Field	Value
Name	Lumesa
Tagline	AI visual infrastructure for e-commerce, generating product imagery for fashion and beauty brands.
Headquarters	New York, North America [Lumesa]
Founded	2021 [Lumesa]
Stage	Seed [Preqin, February 2026]
Business model	SaaS
Industry	E-commerce / Retail
Technology	AI / Machine Learning
Geography	North America
Growth profile	Venture Scale
Founding team	Co-Founders (2), Manvitha Mallela and Malavika Reddy [Forbes, March 2025] [LinkedIn, retrieved 2026]
Funding label	Seed, total disclosed approximately $7.5 million [Bizwomen, October 2026]
Total disclosed	$7.5 million reported total funding, including a $6.0 million seed round led by Work-Bench [Bizwomen, October 2026] [Preqin, February 2026]

Links

Open sources

• Website: https://www.lumesa.ai/

What an Investor Needs First

PUBLIC Lumesa is building AI-generated visual infrastructure for fashion and beauty ecommerce, a category that merits investor attention because the company is selling into a clear cost, speed, and merchandising problem that legacy photoshoots handle poorly [Forbes, March 2024] [Lumesa] [WWD]. Founded in 2021 in New York and previously operating as Flock AI, the company appears to have sharpened its pitch from AI fashion imagery into a broader visual commerce platform, with public materials emphasizing product image generation, personalization, testing, deployment, and virtual try-on for retail brands [Forbes, March 19 2025] [Forbes, March 26 2024] [Lumesa].

The product case rests less on novel model architecture than on workflow fit: Lumesa says brands can generate on-model, product-only, and lifestyle imagery while encoding their own photography guidelines, model preferences, and merchandising rules, which is a more defensible claim than generic image generation if the implementation holds up in production [Lumesa] [Lumesa]. Public evidence for customer traction is still uneven, but there is at least one named brand signal in Anne Klein's beta collaboration on virtual try-on, alongside unattributed claims of production-volume gains and conversion lift that should be treated as directional rather than verified performance benchmarks [WWD] [RocketReach].

The founding team is compact and technically relevant on paper. Public sources identify Manvitha Mallela, also referred to as Vita Mallela, as CEO and founder, and Malavika Reddy as co-founder and CTO; Mallela's public background includes Wharton affiliation, though the stronger read-through for investors is the pair's continued presence across the company's rebrand rather than any over-interpreted credential story [Forbes, March 19 2025] [LinkedIn, retrieved 2026] [The Pitch Podcast, retrieved 2026].

On capitalization, the cleanest confirmed datapoint is a $6.0 million seed round in February 2026 led by Work-Bench, with participation from January Ventures, Outlander VC, The AI Furnace, and Red Bike Capital, while later reporting pegs total funding at $7.5 million and leaves a modest reconciliation gap investors should clarify directly [Preqin, February 2026] [Signalbase] [Bizwomen, October 2026]. The business model is described as SaaS, which fits the product surface, but the more important question over the next 12 to 18 months is whether Lumesa can convert design-led pilots into repeatable enterprise buying behavior without diluting margins through service-heavy onboarding [Forbes, March 2024] [Lumesa].

What deserves close watching now is execution quality: named customer expansion beyond pilots, evidence that virtual try-on and catalog generation can coexist in one coherent platform sale, and whether the company's public traction claims begin to show up in independently verifiable references from brands or investors [WWD] [Bizwomen, October 2026] [Lumesa].

Partially corroborated -- The section relies on a mix of company materials, named media coverage, and one confirmed funding database entry, with several operating claims still company-sourced or only partially corroborated.

Taxonomy Snapshot

Axis	Value
Stage	Seed
Business Model	SaaS
Industry / Vertical	E-commerce / Retail
Technology Type	AI / Machine Learning
Geography	North America
Growth Profile	Venture Scale
Founding Team	Co-Founders (2)
Funding	Seed, total reported funding $7.5M; confirmed $6.0M seed round

Inside the Company

PUBLIC

Lumesa is a New York-based AI visual commerce startup focused on product imagery for fashion and beauty brands, with roots that trace back to its earlier public identity as Flock AI [Lumesa] [Forbes, March 2025]. Public source alignment is reasonably clear on the basic company frame: the business was founded in 2021, operates from New York, and is building software that helps brands generate and manage ecommerce visuals without relying entirely on conventional studio shoots [Lumesa] [Forbes, March 2024].

The company’s founding record is also relatively consistent at the leadership level, even if naming conventions vary across sources. Forbes identifies Manvitha Mallela as CEO of Flock AI, Lumesa’s former name, while LinkedIn identifies Malavika Reddy as co-founder and CTO [Forbes, March 2025] [LinkedIn]. Lumesa’s own website presents the platform around virtual try-on, AI fashion model generation, and visual merchandising workflows for ecommerce teams, which suggests the business has remained centered on retail image production as it moved from Flock AI to the Lumesa brand [Lumesa].

On milestones, the cleanest public chronology starts with the 2021 founding, followed by Lumesa’s inclusion in the NYFTLab 2024 cohort, where Forbes described the company as a visual commerce solution using generative AI for performance-oriented imagery [Forbes, March 2024]. By March 2025, Forbes was still referring to the business as Flock AI in the NYFTLab 2025 cohort context, indicating the rebrand likely sat between those public references or was still settling in market-facing coverage at that point [Forbes, March 2025]. The next clearly documented financing milestone is a $6.0 million seed round reported by Preqin in February 2026, led by Work-Bench [Preqin, February 2026]. A later Bizwomen profile in October 2026 reported $7.5 million in total funding, which is directionally consistent with the seed disclosure but leaves the exact bridge between round size and total capital raised unresolved in public sources [Bizwomen, October 2026].

Partially corroborated -- Core company facts are supported by the company website and Forbes, but legal entity detail is not established here and the total funding figure remains only partially reconciled across Preqin and Bizwomen.

Under the Hood

MIXED

Lumesa is selling a narrower proposition than the broad "AI for retail" label suggests. Public materials describe a system for generating and deploying ecommerce product imagery without conventional shoots, with outputs spanning AI-generated models, product-only images, lifestyle imagery, and virtual try-on [Lumesa, retrieved 2026] [Forbes, March 2024]. The company also frames the product as a visual commerce layer for fashion and beauty brands that need more catalog content with tighter brand control, rather than a general-purpose image tool [Lumesa, retrieved 2026] [Forbes, March 2024].

The more interesting claim is not image generation by itself, but the attempt to codify a brand's visual rules into the workflow. Lumesa says outputs can be styled to a brand's photography guidelines, model preferences, and merchandising rules, and its marketing around digital twins suggests a hybrid model that mixes synthetic talent with licensed likenesses of real people who retain ownership and earn royalties [Lumesa, retrieved 2026]. Press coverage and the company site also place virtual try-on within the product surface, with Anne Klein named in 2026 coverage as a beta partner for that use case [WWD] [Lumesa, retrieved 2026].

There is less public evidence on the underlying stack than on the user-facing workflow, so restraint matters here. No verified engineering documentation in the cited material establishes model architecture, data pipeline design, or deployment method, and claims about training on brand "DNA" remain company-described rather than independently tested [Lumesa, retrieved 2026]. The operating read from public evidence is that Lumesa's differentiation rests on workflow fit and brand-consistency controls more than on a disclosed foundation-model advantage, which is a plausible wedge in fashion ecommerce if customer results hold up in production [Forbes, March 2024] [WWD].

Claim stands unchecked -- This section relies materially on company website claims, with partial support from Forbes and WWD on product scope and use cases.

Market Research

PUBLIC

The market matters now because apparel and beauty brands are under pressure to produce more product imagery across more channels, while the traditional photo production process remains slow, sample-heavy, and expensive [Forbes, March 2024] [LinkedIn, retrieved 2026].

Lumesa sits at the intersection of several better-documented markets rather than a single clean category. The company is described publicly as a visual commerce and generative AI infrastructure product for ecommerce imagery [Forbes, March 2024] [Lumesa]. No named third-party market report in the supplied record gives a direct TAM, SAM, or SOM for AI fashion imagery, so the cleaner approach is to frame Lumesa against analogous markets: ecommerce enablement software, retail visual merchandising technology, and generative AI creative tooling (analogous market, source). That matters because the budget owner is likely not a standalone "AI imagery" line item, but a mix of ecommerce operations, creative production, merchandising, and performance marketing spend [Forbes, March 2024] [Lumesa].

The demand signal in the public record is less about category size than about workflow pain. Lumesa's pitch is built around replacing or reducing samples, models, studios, and photoshoots for catalog and campaign imagery [Lumesa] [Forbes, March 2024]. One cost benchmark surfaced in a founder-linked LinkedIn profile states that showing one garment in one color on 20 bodies can cost $19,000 to $38,000 in a traditional shoot [LinkedIn, retrieved 2026]. That is a weakly verified number, but directionally it helps explain why buyers would test software that can generate on-model, product-only, lifestyle, and virtual try-on assets from existing product data rather than schedule new shoots [Lumesa] [Forbes, March 2024].

Adjacent markets broaden the opportunity but also clarify where substitution risk sits. Lumesa can be read as part of the product information and merchandising stack, part of the digital asset creation stack, and part of the conversion-optimization stack when imagery is tested for performance [Lumesa] [Forbes, March 2024]. Substitute solutions include conventional photo studios, CGI and 3D visualization workflows, virtual try-on tools, and internal creative teams using general-purpose image models. The market question is therefore not only whether brands want AI-generated imagery, but whether a vertical product trained to a brand's own photography guidelines and merchandising rules can outperform those substitutes on approval rates, speed, and conversion impact [Lumesa].

Regulatory and macro forces are mixed. On the supportive side, retailers have strong incentives to lower content production cost and shorten go-to-market cycles, especially in categories with frequent assortment turns and high SKU counts [Forbes, March 2024]. On the constraining side, Lumesa's own product framing around licensed digital twins and royalty-bearing use of real talent points to a rights-management layer that is becoming more important as brands face scrutiny over consent, likeness usage, and disclosure in AI-generated media [Lumesa]. That does not create a direct barrier unique to Lumesa, but it does mean adoption in fashion and beauty may favor vendors that can document provenance, permissions, and brand-control workflows rather than simply generate attractive images [Lumesa] [WWD].

Market lens	Public sizing or evidence	Relevance to Lumesa
Traditional apparel image production economics	One founder-linked public benchmark cites $19,000 to $38,000 to show one garment in one color on 20 bodies in a traditional shoot [LinkedIn, retrieved 2026]	Supports the cost-savings wedge if the benchmark is representative
Retail tech and visual commerce adoption	Lumesa was included in NYFTLab's 2024 cohort of women-led B2B retail tech startups, with Forbes describing its product as a visual commerce solution for performance-optimized imagery using generative AI [Forbes, March 2024]	Suggests the product maps to an active retail software buying category
Enterprise experimentation in virtual try-on	Anne Klein launched a beta collaboration with Lumesa to test AI virtual try-on, according to WWD [WWD]	Indicates demand extends beyond catalog generation into shopper-facing conversion tools

The table points to a market that is easier to validate through budget pain and adjacent software adoption than through a clean top-down category definition. For investors, that usually shifts diligence toward buyer behavior, replacement dynamics, and procurement friction rather than headline TAM slides.

Partially corroborated -- This section relies on named public sources such as Forbes, WWD, LinkedIn, and Lumesa's website, but it does not have third-party market sizing data for a direct TAM/SAM/SOM view.

Competition and Substitutes

MIXED Lumesa is positioned against a narrow set of AI imagery vendors for commerce, and against a much broader substitute set that still includes traditional photo production, internal creative teams, and retailer experimentation with point tools [Lumesa, retrieved 2026] [Forbes, March 2024].

Company	Positioning	Stage / Funding	Notable Differentiator	Source
Lumesa	AI visual infrastructure for ecommerce, focused on generating and deploying product imagery for fashion and beauty brands	Seed, $6.0M seed round in February 2026; $7.5M total funding reported separately	Emphasizes brand-specific imagery generation, including outputs styled to photography guidelines, model preferences, and merchandising rules	[Preqin, February 2026] [Bizwomen, October 2026] [Lumesa, retrieved 2026]
Flair AI	AI design and branded content tooling for ecommerce brands	Not established in the provided public source set	Named competitor in structured research	[Structured Facts]

The competitive map breaks into three practical layers. Second are adjacent software providers in visual commerce, including virtual try-on and 3D visualization tools, which compete for the same budget line even when their product architecture differs [WWD] [Lumesa, retrieved 2026]. Third are substitutes that are not software startups at all, namely conventional shoots, agency production, and in-house creative operations. That last layer still matters because Lumesa's pitch is not simply better software, it is a rewrite of how fashion and beauty content gets produced [LinkedIn, retrieved 2026] [Forbes, March 2024].

Lumesa's edge, on the public record, appears to rest less on a model layer than on workflow fit and brand control. The company says its system generates imagery according to a brand's own photography guidelines, styling rules, pose libraries, model preferences, and merchandising standards, which is a more specific claim than generic text-to-image creation [Lumesa, retrieved 2026]. If that claim holds in practice, the advantage is meaningful because large retailers do not buy images in the abstract, they buy consistency across channels, seasons, and SKU volumes. The same evidence also points to the main durability question: this edge is only durable if those brand-specific rules and approvals compound into a hard-to-replicate data asset. Public sources do not yet show whether Lumesa owns a distribution channel, exclusive customer data, or a network effect that would make switching materially harder over time [Lumesa, retrieved 2026] [Forbes, March 2024].

The company's main exposure is that several better-capitalized or better-distributed alternatives can approach the same buyer from different angles. At the same time, adjacent vendors in virtual try-on or 3D commerce can win budget if the retailer's primary problem is fit confidence or interactive merchandising rather than catalog throughput [WWD] [Lumesa, retrieved 2026]. Lumesa also does not appear, from public evidence here, to own the commerce platform, ad channel, or retailer operating system through which images are ultimately distributed. That leaves it selling into a crowded application layer where procurement can remain experimental.

Over the next 18 months, the most plausible competitive outcome is a split market. Lumesa is a winner if enterprise fashion and beauty brands keep prioritizing brand-consistent asset production over general-purpose image generation, because that favors vendors that can encode house rules and production workflows rather than simply render attractive outputs [Lumesa, retrieved 2026]. The most exposed player under that scenario would be any vendor, including Lumesa, that cannot show named customer references and repeatable deployment evidence at scale. Publicly, Lumesa has one named beta-style collaboration with Anne Klein and broader but mostly unnamed customer claims, which leaves the competitive proof point narrower than the category ambition [WWD] [Bizwomen, October 2026].

Reasoned from indirect evidence -- This section uses public sources, but the competitive set is thinly disclosed. Lumesa's positioning is supported by company materials and Forbes, while the named competitor list is limited and several operational claims remain company-led or structured-research only.

Opportunity

Open sources

If Lumesa executes, the prize is not a single creative tool vendor outcome, but a chance to become the operating layer fashion and beauty brands use to produce, test, and refresh product imagery at ecommerce scale, a budget line that sits close to revenue rather than brand experimentation [Lumesa; Forbes, March 2024].

The headline opportunity is straightforward: own the workflow that turns a SKU into commercially usable imagery across catalog, lifestyle, and virtual try-on surfaces, without the time and cost burden of repeated physical shoots [Lumesa; Forbes, March 2024]. That outcome looks reachable, not merely aspirational, because the company is already positioned around production use cases rather than concept demos. Public materials describe generated models, product-only imagery, lifestyle imagery, and virtual try-on in one stack, and third-party coverage ties the product to performance-oriented visual commerce rather than pure design software [Lumesa; Forbes, March 2024]. The most credible wedge is fashion and beauty, where merchandising velocity matters and visual inconsistency directly affects conversion, returns, and content throughput. The reported Anne Klein beta adds one named brand reference point for market receptivity, even if it does not yet establish scaled deployment economics [WWD; Breaking Tech].

The growth paths are still narrow enough to name. That is useful, because this business likely scales through a small number of commercial motions rather than one broad consumer breakout.

Scenario	What happens	Catalyst	Why it's plausible
Enterprise catalog system of record	Lumesa becomes the default imagery workflow for large apparel and accessories groups, replacing a material share of sample-driven catalog production with AI-generated assets	A few visible enterprise references convert beta usage into standardized rollouts across multiple brands or business units [WWD; Bizwomen, October 2026]	The product is already framed around brand-specific rules and merchandising consistency, which is the requirement for enterprise adoption, not just image generation quality [Lumesa; Forbes, March 2024]
Embedded visual commerce layer	The company sells into ecommerce stacks as an embedded service for imagery generation, testing, and virtual try-on rather than a stand-alone seat-based tool	A category-tipping integration or partnership with a major retailer, platform, or agency network that needs imagery at volume [WWD; Forbes, March 2025]	Public positioning as "visual infrastructure" suggests management is aiming at workflow embedment, and the Anne Klein beta shows at least one live test tied to the shopping journey [Lumesa; WWD]
AI model network with rights management	Lumesa turns digital twins and licensed AI talent into a differentiated supply layer, giving brands a compliant way to scale inclusive imagery while compensating talent	Wider adoption of licensed digital twins as brands seek lower-friction content production with fewer legal and brand-safety questions [Lumesa; FashionUnited, October 2026]	The company already markets custom AI models and licensed digital twins with royalty structures, which could matter if the market shifts from generic AI faces toward auditable, rights-cleared assets [Lumesa]

What compounding looks like here is a brand-control flywheel. If a retailer feeds Lumesa more of its photography guidelines, styling rules, pose libraries, and merchandising preferences, the output becomes more usable in production, which should lower review time and make the next campaign easier to approve [Lumesa]. Better approval rates can widen use from a single content team to multiple categories, and each expansion raises switching costs because the system is no longer just generating images, it is encoding visual standards.

There is also a commerce-performance flywheel if the product can keep tying imagery changes to measurable results. Company-linked claims reference a 3 percent conversion lift at a global luxury accessories holding company and doubled production volume at a large apparel group, though those claims remain unnamed and should be treated cautiously [RocketReach; Lumesa]. Even so, the direction of the value proposition is clear: if buyers see faster asset creation and better onsite performance in the same workflow, budget can move from creative experimentation to core merchandising spend. The reported presence of more than 100 brands on or trusting the platform would support that thesis if later corroborated, but today it remains company-linked rather than independently verified [Lumesa].

The size of the win is best framed against software infrastructure economics rather than agency revenue. No public market comparable is provided in the source set, and no verified market-sizing report is captured here, so precision would be false. A reasonable upside case is that, if Lumesa became a standard system for enterprise visual merchandising across fashion and beauty, it could support a venture-scale software outcome in the hundreds of millions of dollars of enterprise value, or more, because it would sit in a recurring workflow tied to SKU launches, seasonal refreshes, and conversion optimization (scenario, not a forecast). That scenario rests on one key public premise: the company is trying to replace a recurring operational process, not sell one-off creative output, and the economics of traditional shoots appear high enough to make substitution budgets meaningful when quality is acceptable [LinkedIn, retrieved 2026; Forbes, March 2024].

Partially corroborated -- This section relies on a mix of company materials and limited third-party coverage, with some commercial outcome claims supported only by company-linked or secondary sources such as Lumesa, RocketReach, WWD, Breaking Tech, and Forbes.

Sources

Open sources

1. [Lumesa] Lumesa ai | Virtual try-on for ecommerce, AI fashion model generator, ai tool | https://www.lumesa.ai/

2. [Preqin, February 2026] Lumesa (New York) 2026 Company Profile: Valuation, Funding & Investors | https://pitchbook.com/profiles/company/536040-73

3. [Forbes, March 2025] The Future Of Fashion Tech: NYFTLab Announces Its 2025 Cohort | https://www.forbes.com/sites/angelachan/2025/03/19/the-future-of-fashion-tech-nyftlab-announces-its-2025-cohort/

4. [LinkedIn, retrieved 2026] Malavika Reddy - Co-founder & CTO at Flock AI | https://www.linkedin.com/in/malavikareddy/

5. [Bizwomen, October 2026] Lumesa CEO Vita Mallela raises $7.5M for retail AI startup - Bizwomen | https://www.bizjournals.com/denver/bizwomen/news/profiles-strategies/2026/10/vita-mallela-lumesa.html

6. [Forbes, March 19 2025] The Future Of Fashion Tech: NYFTLab Announces Its 2025 Cohort | https://www.forbes.com/sites/angelachan/2025/03/19/the-future-of-fashion-tech-nyftlab-announces-its-2025-cohort/

7. [Forbes, March 26 2024] NYFTLab Unveils The 2024 Cohort Of Six Women-Led B2B Retail Tech Startups In Partnership With Leading Retailers And Brands | https://www.forbes.com/sites/angelachan/2024/03/26/nyftlab-unveils-the-2024-cohort-of-six-women-led-b2b-retail-tech-startups-in-partnership-with-leading-retailers-and-brands/

8. [WWD] Anne Klein Taps Lumesa For A More Immersive Virtual Try-On Experience | https://wwd.com/sourcing-journal/industry-news/anne-klein-lumesa-virtual-try-on-fitting-room-ai-1239202760/

9. [The Pitch Podcast, retrieved 2026] Best B2B SaaS Startup Pitches | The Pitch Podcast | https://www.thepitch.show/categories/b2b-saas-startup-pitches

10. [Signalbase] Lumesa Secures $6.0M | https://www.trysignalbase.com/news/funding/lumesa-secures-6m

11. [LinkedIn, retrieved 2026] Daphne Greig - Northcott silk inc. | LinkedIn | https://www.linkedin.com/in/daphnegreig/