How a Small AI Startup Won Over Big Retailers by Learning to Work Their Way

How does a small startup persuade billion-dollar retailers to adopt an unfamiliar technology? Lumesa co-founder Vita Mallela used her experience at Bloomingdale’s and Walmart to understand the challenges facing fashion brands. Instead of forcing customers to change their operations, Lumesa began with manual services, learned how its clients worked, and developed technology around their needs. The company’s experience offers eight practical lessons in customer acquisition, building trust, enterprise sales, and scaling a business.

Key Takeaways

• Experience inside large organizations can uncover opportunities. While working at major retailers, Vita Mallela recognized an expensive, recurring problem in fashion imagery production.
• Winning enterprise customers requires more than an innovative product. Lumesa gained early trust by working within brands’ existing processes rather than forcing them to adopt unfamiliar systems.
• Manual work can be a competitive advantage during the early stages. Direct customer collaboration helped the founders understand what their eventual technology needed to automate.
• Corporate customers need reassurance about what could go wrong. Protecting brand identity, product accuracy, and model likenesses became important parts of Lumesa’s approach.
• An initial engagement can lead to broader opportunities. Lumesa’s work supporting imagery for 1,500 SKUs expanded into additional marketing and campaign applications, according to the company.
• Technology needs a clear economic justification. Large customers evaluate solutions in terms of costs, production efficiency, risk, and operational results.
• New products should build on existing customer needs. Lumesa’s expansion into virtual try-on reflects its original focus on helping retailers present their merchandise more effectively.
• Small businesses can compete through responsiveness and specialized knowledge. Direct customer relationships and flexibility can help smaller suppliers establish credibility with much larger organizations.

Before Lumesa raised $7.5 million, its founders worked with fashion brands via email, Google Drive, and painstaking manual processes. Their approach offers an important lesson for small businesses hoping to sell to large companies: understanding how customers work can matter just as much as what your product can do.

A small technology company trying to sell to a billion-dollar retailer faces an obvious disadvantage. The retailer already has established suppliers, internal teams, purchasing procedures, and years of investment in its systems. Any new vendor needs to offer something compelling enough to justify taking a chance on an unfamiliar business.

Having an innovative product is rarely sufficient. Large companies also want confidence that a new solution will meet their standards, work with existing processes, and avoid creating problems for the people responsible for implementing it.

Vita Mallela understood those concerns long before she became an entrepreneur.

>> Related Reading: In our companion feature on PowerHomeBiz.com, we examine how Vita Mallela identified the opportunity behind Lumesa, built the company with co-founder Malavika Reddy, and raised $7.5 million in funding.

During her years in merchandising and creative operations at Bloomingdale’s and Jet.com/Walmart, she saw how much work went into producing the photographs used to sell clothing online. She also understood the pressure on retail teams to create more imagery, represent more customers, and support increasingly large product catalogs without allowing production costs to spiral.

By early 2024, advances in generative artificial intelligence suggested that some of this work could be done differently. Mallela recognized the opportunity and co-founded Flock AI, now known as Lumesa, with Malavika Reddy.

The company develops AI technology that allows fashion and retail brands to create product imagery while maintaining their visual identity and creative standards. In February 2026, the company announced a $6 million seed round led by Work-Bench, bringing its total funding to $7.5 million.

What makes Lumesa particularly interesting for Learning From Big Boys is not simply its technology or its funding. It is how a young company approached some of the most demanding customers in retail.

In an interview conducted for our sister publication, PowerHomeBiz.com, Mallela described an early business that was far more manual than most people might expect from an AI startup. Rather than building an elaborate platform and expecting brands to adapt, the founders worked directly with customers, learned their requirements, and gradually developed technology around what those customers needed.

The experience offers several useful lessons for entrepreneurs trying to enter markets dominated by much larger companies.

Lesson 1: Working Inside a Big Company Can Reveal Opportunities Outsiders Miss

One advantage of large companies is access to people who understand complex industries from the inside. Those employees know how decisions are made, where money goes, and which operational problems persist despite repeated attempts to solve them.

Yet that familiarity can also create blind spots. A process that has existed for years may eventually be accepted as the normal cost of doing business, even when it is inefficient.

Mallela encountered just such a problem in fashion photography.

Producing images for ecommerce involves coordinating models, studios, stylists, photographers, merchandise samples, and retouching. When brands want to show clothing on additional body types or models, they often need to repeat parts of that production process.

As Mallela explained, retailers were not necessarily opposed to creating more representative imagery. The economics made it difficult.

“I was seeing the same gap from both sides,” she said, describing the limited range of models featured in fashion imagery and the operational constraints that contributed to that problem.

Because she had worked inside major retailers, Mallela could see the relationship between an apparently creative issue and the underlying business process. The problem was not simply a shortage of photographs. It was a production system that made creating more photographs expensive and difficult to scale.

That distinction gave her a clearer understanding of what a commercially useful solution would need to accomplish.

The U.S. Small Business Administration’s guidance on market research and competitive analysis encourages entrepreneurs to examine customer demand, existing alternatives, and market conditions before committing resources to a business idea. Understanding why customers remain dissatisfied with those alternatives can be especially valuable.

For entrepreneurs, the opportunity is not limited to industries experiencing major technological disruption. Similar problems exist in construction, healthcare administration, professional services, manufacturing, logistics, and countless other sectors.

An employee who repeatedly encounters delays in obtaining approvals, coordinating vendors, or managing information may be observing an opportunity for a new service or product.

The challenge is determining whether the problem is important enough that customers would pay to solve it.

Our earlier Learning From Big Boys feature, The Problems Everyone Else Has Learned to Live With: Joanne Frederick’s 30-Year Business Lesson, explores a similar principle. Valuable business opportunities often emerge from problems organizations have stopped questioning because they have been part of everyday operations for so long.

What small businesses can learn: Before searching for an entirely new idea, examine the processes that people in your industry regularly complain about. The most promising opportunities may involve work everyone assumes is difficult.

Lesson 2: Large Customers Do Not Want Innovation That Creates More Work

Startups often believe a superior product is enough to persuade customers to switch. But for a large organization, adopting something new can create substantial work even when the technology promises long-term savings.

Employees may need training. Purchasing departments must review contracts. Managers need to approve changes. Existing workflows may require modifications, and someone must take responsibility if implementation does not go smoothly.

Lumesa’s founders approached the problem by asking customers to make as few changes as possible.

“The earliest version was very hands-on,” Mallela recalled. “There was no automated brand DNA and not even a full dashboard yet.”

The company worked with fashion brands via email and Google Drive, reviewing existing imagery and manually identifying the visual characteristics to preserve.

The founders studied how garments should appear, which poses suited a brand, and what lighting and styling standards mattered to its creative teams. Customers reviewed the work throughout production rather than being handed finished images without explanation.

The process was time-consuming, but it reduced an important source of resistance.

“We weren’t asking brands to hand their identity over to a black box,” Mallela explained.

Instead, the founders asked customers what information they were comfortable sharing, what aspects of their brand mattered most, how they preferred to review work, and what deadlines needed to be met.

For an established retailer, that made the new technology less disruptive. Creative teams retained influence over the outcome, while Lumesa learned how to operate within their expectations.

This approach is applicable to almost any small company selling to a larger organization.

A consulting firm offering a new reporting service might initially deliver results in the format the customer’s managers already use rather than insisting on an unfamiliar dashboard. A software startup might integrate with the customer’s existing systems instead of demanding a complete operational overhaul. A manufacturer might adjust packaging, delivery, or documentation to fit a retailer’s established procedures.

The goal is not to agree to every customer request. It is to recognize that adoption has a cost, and reducing that cost can make an otherwise unfamiliar solution easier to accept.

What small businesses can learn: When approaching a major customer, don’t focus exclusively on your product’s benefits. Consider what the customer will have to change to use it and find ways to make that transition easier.

Lesson 3: Your First Customers Can Help You Figure Out What to Automate

At first glance, Lumesa’s early reliance on manual processes might seem unusual for a company built around artificial intelligence.

But those processes served an important purpose. They let the founders observe customer requirements before committing to a specific software design.

Every conversation about lighting, product accuracy, styling, or approvals helped them identify the tasks that repeatedly mattered across customers. Those recurring requirements eventually became the foundation of the company’s Brand DNA system.

“That manual process taught us exactly what to automate,” Mallela said.

Today, the company says Brand DNA incorporates more than 200 attributes of a retailer’s visual identity. These include photography rules, preferred poses, styling standards, and other details intended to help generated images match a brand’s established appearance.

The company’s technology grew from an understanding of what customers were trying to accomplish, not from a collection of features developed in isolation.

This is a useful distinction for entrepreneurs investing in automation.

A business owner might see a competitor using an expensive customer relationship management system and assume that purchasing similar software will improve sales. A service provider might decide to automate every part of its operations before learning which tasks require customization or personal attention.

Automation can improve efficiency, but it can also make an ineffective process faster without addressing the underlying problem.

Lumesa’s experience suggests a different sequence: perform the work, observe where time and resources are consumed, identify recurring patterns, and then automate the activities that can be standardized without compromising quality.

There is a trade-off, of course. A highly manual approach becomes expensive if the founder cannot eventually transfer knowledge into systems, employee procedures, or repeatable methods.

The objective is not to remain dependent on manual work indefinitely. It is to use that work to learn what the business needs before investing heavily in automation.

Our profile of Kecia Hielscher and the big-company lessons she brought to Formulary 55 explores a related idea. Experience inside established organizations can teach entrepreneurs the value of consistent processes, but smaller companies need to adapt those practices to their own size and resources.

What small businesses can learn: Don’t rush to automate a process you don’t yet understand. Early hands-on work can help identify the standards, exceptions, and customer preferences that your eventual systems must accommodate.

Lesson 4: Winning Corporate Trust Requires Understanding What Customers Are Afraid of Losing

For a fashion brand, adopting AI-generated imagery creates risks beyond the possibility that the technology might not work.

The brand has spent years developing a recognizable visual identity. Its photographs communicate product quality, positioning, style, and expectations. Customers rely on those images to decide whether to buy.

If AI-generated content misrepresents the clothing or produces visuals that feel inconsistent with the brand, the savings from producing the imagery may not justify the consequences.

Mallela said Lumesa faced questions about product accuracy, creative control, consumer trust, and using models’ likenesses.

The company responded by building safeguards around those concerns.

Its Brand DNA system is intended to preserve each retailer’s established visual standards. For human likenesses, Lumesa says it uses licensed digital twins of real models who consent to the use of their images and are compensated.

The company also advocates clear disclosure of AI-generated content.

“The brands that get this right will treat a person’s likeness the way they treat their own logo: licensed, protected, and never used without permission,” Mallela said.

For larger companies, these considerations can determine whether a project receives approval. Even a technically impressive solution may be rejected if it creates unacceptable legal, operational, or reputational risks.

The Federal Trade Commission’s advertising and marketing guidance reinforces the importance of truthful, substantiated claims and avoiding misleading commercial representations. The same obligations remain relevant when marketing content is created with AI.

Small businesses can learn from Lumesa’s approach by recognizing that customers may resist a new product because they fear losing something, not because they don’t understand its benefits.

An accounting firm considering new software may worry about data security. A manufacturer considering a new supplier may worry about delivery reliability. A company outsourcing customer support may worry about losing control of its customer relationships.

A persuasive sales presentation should address those concerns with evidence, transparent processes, and reasonable safeguards.

What small businesses can learn: Find out what your prospective customers cannot afford to compromise. Demonstrating that you understand and protect those priorities can be more convincing than simply promising lower prices or faster service.

Lesson 5: A Small Vendor Can Become More Valuable by Solving a Bigger Operational Problem

Lumesa’s growth also illustrates how a business relationship can expand when a supplier shows it can solve a recurring operational problem.

Mallela described a project involving a multi-brand apparel holding company valued at more than $1 billion. The customer needed far more product imagery than its traditional studio could produce within existing time and budget constraints.

According to Mallela, Lumesa supported imagery for 1,500 stock-keeping units, or SKUs, using AI technology configured around the company’s visual standards.

The relationship later expanded into marketing and campaign work across multiple brands within the holding company’s portfolio.

The client was not named, and the underlying project metrics were not independently provided. Nevertheless, the example shows the kind of expansion opportunity that can develop when an initial service proves useful to a large customer.

Lumesa was initially addressing a production bottleneck. Once the customer became comfortable with the technology and its output, it could use it for additional applications.

There is a broader business-development lesson here.

A small supplier does not necessarily need to enter a large organization by selling its biggest or most expensive service. In many cases, a limited engagement with measurable objectives is easier for the customer to approve and evaluate.

A marketing agency might start by improving one underperforming campaign before taking on a broader account. An IT consultant might solve a recurring reporting problem before proposing a larger integration project. A specialized manufacturer might begin with a limited product line before expanding its supply agreement.

The first engagement gives the supplier an opportunity to demonstrate reliability, understand internal decision-making, and build relationships with people who influence purchasing decisions.

But expansion must be earned. A customer that is satisfied with one service will not automatically buy another, especially if the new offering involves different budgets or decision-makers.

This connects with an observation in our Learning From Big Boys interview with Joanne Frederick: an introduction may create the first opportunity, but subsequent business depends on how well a company delivers.

What small businesses can learn: Consider how a limited first engagement could demonstrate value and create opportunities for additional work. Concentrate on delivering a strong initial result rather than trying to sell everything at once.

Lesson 6: Customers Pay for Business Results, Not Technology Alone

Artificial intelligence has created considerable excitement, but customers ultimately have to justify spending money on it.

For an enterprise retailer, the important questions are practical. Can the technology reduce content-production costs? Will it help creative teams meet deadlines? Can the retailer produce more product imagery without hiring additional staff or arranging more photoshoots?

Mallela said brands using systems like Lumesa’s can save 70% to 90% compared with traditional physical photography while producing more images on the same budget.

These are company-reported estimates, not independently verified industry benchmarks. Actual savings would depend on photography requirements, licensing costs, revisions, quality-control work, and implementation expenses.

Nevertheless, the figures demonstrate the economic argument the company is making to prospective customers.

Lumesa is not merely selling the ability to create pictures using AI. It is selling a potential improvement in the economics of producing commercial imagery.

That distinction becomes especially important when approaching major corporate customers. Different departments may evaluate the same product differently. Creative teams may care about image quality and brand consistency. Operations managers may focus on production time. Finance departments may want evidence of reduced expenses or improved return on investment.

A successful vendor needs to understand these priorities and communicate the relevant benefits without overstating what the product can deliver.

This is also why the company’s evolution beyond basic AI-generated imagery matters. Lumesa has expanded into virtual try-on and broader visual-commerce applications, giving brands more ways to present products to shoppers.

Mallela expects virtual try-on to become increasingly important as large platforms such as Google and Walmart bring the technology into mainstream shopping. But she argues that retailers will need more than a generic capability. The experience must preserve the brand’s identity and fit naturally into the customer’s shopping journey.

For smaller businesses, the parallel is straightforward. Customers generally care less about a product’s sophistication than about whether it solves an expensive, frustrating, or time-consuming problem.

Evaluate a new scheduling application by how much time it saves and whether it reduces missed appointments. Judge a marketing service by the business outcomes it helps produce, not simply the number of reports it delivers.

What small businesses can learn: Translate product features into results customers understand. If a product promises savings or improved performance, explain how those benefits are measured and what conditions affect the outcome.

Lesson 7: The Best Product Expansion Often Comes From the Original Customer Problem

Technology companies face constant pressure to introduce new features, especially in fast-moving markets like artificial intelligence.

But expanding into every new capability can weaken a company’s focus. Development resources are limited, and each additional feature introduces implementation, support, and maintenance requirements.

Lumesa’s expansion shows how a company can pursue new opportunities while staying connected to the problem that created the business.

The company began by helping fashion brands create more product imagery. Its technology focused on accurate garment representation, brand-specific visual standards, and greater flexibility in how merchandise can be presented.

Virtual try-on extends those capabilities into the shopping experience itself. Instead of displaying a garment on several selected models, retailers can potentially help customers visualize how an item might look on them.

The applications differ, but they share underlying requirements for accurate visual representation, product information, and customer confidence.

Mallela sees the next challenge as connecting these capabilities rather than treating each one as an isolated technology feature.

For entrepreneurs, this distinction can help separate sensible expansion from unnecessary diversification.

A company that builds accounting software for restaurants may see an opportunity to develop inventory reporting because the same customers need both. A business specializing in commercial cleaning may identify demand for related maintenance services because it already understands clients’ facilities and operating schedules.

These extensions make sense when they build on existing knowledge, relationships, and operational capabilities. They become harder when the company enters an unfamiliar market solely because the opportunity looks attractive.

Our analysis of how Rugs Direct has survived repeated changes in ecommerce explores a related principle. Technologies and shopping habits evolve, but the underlying customer need can remain remarkably consistent.

For Rugs Direct, helping shoppers feel confident buying products that are difficult to visualize remains important despite changes in digital commerce. Lumesa addresses a related challenge on the retailer side, helping brands present products accurately and effectively across different visual formats.

What small businesses can learn: When considering a new product or service, ask whether it strengthens your ability to solve the original customer problem. Growth is more manageable when each expansion builds on capabilities the company already understands.

Lesson 8: Small Companies Can Compete by Being Easier to Work With

Large companies have advantages that small businesses cannot easily reproduce. They have greater financial resources, established customer relationships, larger teams, and often more sophisticated technology.

But size can also introduce complexity. Decisions may require multiple approvals, product changes can involve lengthy development cycles, and customers may struggle to obtain personalized attention.

A smaller company can sometimes use its limited size to its advantage.

Lumesa’s early founders were directly involved in customer conversations, image reviews, and creative decisions. They did not have a fully developed product or extensive internal infrastructure, but they could work closely with customers and adjust their processes based on feedback.

That proximity provided information that might have been harder to obtain through a more standardized customer-service arrangement.

It also helped build relationships with companies understandably cautious about adopting generative AI in commercially sensitive applications.

This does not mean manual service is inherently superior to a well-developed system. A business that depends on its founders personally managing every customer detail will eventually hit growth limits.

The advantage comes from using direct customer access to learn quickly, make informed decisions, and develop repeatable practices without losing the responsiveness that attracted customers in the first place.

That balance is one of the central challenges of scaling a small company.

An entrepreneur may initially win business because customers enjoy dealing directly with the owner. As the company grows, employees, documented processes, and systems must support those relationships and deliver consistent results.

If growth causes service quality to deteriorate, the business may lose the very advantage that helped it compete.

What small businesses can learn: Being smaller can make it easier to listen, respond, and customize. Use those strengths deliberately, but build systems that let the company maintain its standards as the customer base expands.

What Lumesa’s Experience Says About Selling to Big Companies

Lumesa’s story reminds us that innovation and customer adoption are two separate challenges.

Developing technology that generates realistic fashion imagery is one achievement. Convincing major retailers to trust that technology with their products, visual identities, and customer relationships is another.

Mallela’s years at Bloomingdale’s and Walmart helped her understand both sides of that challenge. She knew the production problems retailers faced, but she also understood why those companies couldn’t simply abandon established processes whenever a promising new technology appeared.

That knowledge shaped how Lumesa approached its first customers. The founders started by learning how brands operated, involving their teams in the production process, and building technology around requirements they encountered repeatedly.

The company’s subsequent funding and expansion show that the approach has attracted investor interest and enterprise customers. They do not, by themselves, establish long-term profitability or prove that the business model will succeed under every market condition.

For small businesses, however, the underlying lessons are practical and applicable across industries.

Large customers don’t necessarily want the most elaborate solution. They want suppliers who understand their requirements, deliver reliable results, and make their work easier without creating unacceptable risks.

A young company may not match an established competitor’s resources, but it can often compete through preparation, specialized knowledge, flexibility, and close attention to customer needs.

Lumesa’s founders started by working through email and Google Drive to understand what fashion brands needed. The technology became more sophisticated over time, but the company started with something much more familiar: listening closely to customers and doing the difficult work required to earn their confidence.

For entrepreneurs hoping to win business from larger companies, that may be the most useful lesson of all.

Frequently Asked Questions