Jehan Hamedi and the picture worth a second look

Before Jehan Hamedi began teaching computers to assess pictures, he watched his parents try to find customers. They ran a manufacturing business in Wisconsin. Getting new leads could be difficult. Their son eventually turned that ordinary, stubborn business problem into a question: could the people a company wanted to reach supply the clues to reaching them?

It is a useful place to begin his story because the problem is so unglamorous. A business can make something worth buying and still struggle to attract a buyer. Hamedi's early answer was software that translated public consumer activity into marketing advice. The distance between a factory looking for leads and a photograph looking for a click is shorter than it first appears.

He moved east for college and graduated from Boston University's Questrom School of Business in 2012, studying business administration, marketing, and business law. Boston became home. In his account of those years, it is also where he met his wife. His career would keep returning to the same intersection: what people reveal about their preferences, and what businesses do with that knowledge.

First, listen to the crowd

At Crimson Hexagon, Hamedi worked on making social data useful to companies. In 2012, he was global market development manager, responsible for translating the software's capabilities into propositions and strategies for different industries and international markets. There was a commercial education inside the technical one: a company needs to understand why an analysis matters before it will pay for it.

An unusual assignment arrived one evening when his CEO asked what they should show Prince William and Princess Catherine. Hamedi recalls assembling an analysis of Canadian social-media discussion about the couple, which they personally reviewed. His route into a royal briefing ran through a dataset. For a young analytics executive, there are less memorable ways to spend an evening.

His work brought him into contact with large brands, technology companies, and entertainment businesses. These were organizations with substantial consumer information and the resources to act on it. Hamedi became interested in the advantage that information conferred. The question was how to make that advantage available through a product, with guidance a marketer could actually use.

Adhark's Cuckoo, announced in June 2016, gave that idea a practical form. A user could ask marketing questions about a target audience and receive recommendations drawn from social activity. The launch described more than 20 algorithms producing suggestions across content, influencers, design, and community management. It offered a free starting point to businesses or creators with a Twitter handle. A bird's name carried a fairly earnest ambition: help people decide what to do next.

A color could be a clue

Cuckoo looked at groups of people rather than simply counting reactions to a campaign hashtag. A retailer could investigate the people engaging with another brand and learn about their interests. Hamedi was already discussing image analysis as a further opportunity. The colors that attracted an audience might contain information a text-only report would miss.

There was also a lesson in the software's own presentation. During Adhark's public beta, his team found that valuable information still needed a clear route through the interface. A useful answer buried in an awkward workflow asks the customer to finish the product's work. He described improving that experience and repeatedly asking customers for feedback.

That makes his later interest in images less of a sudden conversion than a continuation. He had been dealing with two related problems: how to discover what an audience likes, and how to present information so someone can act on it. A picture on a product page compresses both into a small rectangle. Every decision about that rectangle has to survive an encounter with somebody who has other things to do.

Giving a photograph a job description

Vizit emerged publicly in 2020, and its Visual Brand Performance platform launched globally in January 2021. Hamedi described a browser-based service that could connect to customers' content libraries and asset workflows. Its early features included attention maps and visual sentiment maps, overlays intended to show where an audience would look and which portions of an image helped or hurt its effectiveness.

The premise gave a familiar creative argument a different place to go. A team could ask how a particular audience might respond to an image before committing to it. Hamedi's early ambition was broad: bring this kind of evaluation into the archives of existing images as well as the flow of new work. Brands including Harley-Davidson, Reebok, and Master Lock were among the customers he named in 2021.

The path to a software platform was practical rather than immaculate. In a founder conversation, he described early engagements that resembled technology-enabled consulting. The data and intellectual property existed before the full interface did. Working closely with customers helped identify where the technology produced results worth paying for. A company can spend a long time polishing a dashboard. A customer with a specific problem provides a more useful deadline.

The faucet has an audience

Hamedi's explanations become especially clear when the subject is a faucet. Discussing work with Moen, he described questions about kitchen settings, text placement, and the position of the fixture itself. Camera angle, background, and composition became things to evaluate against the intended audience. He reported a 20 percent conversion gain after optimized content went live. That is a result he described for that engagement, rather than a promise for every photograph.

The underlying approach uses synthetic audiences, which Vizit calls Audience Lenses, to predict responses to imagery. The output is audience-specific. A formula such as making every background pink would miss the relationship among the product, the category, and the person shopping. Hamedi's work puts pressure on the comfortable idea that a universal set of aesthetic rules will settle every merchandising decision.

There is a comic indignity to asking whether a faucet has been photographed from its good side. There is also a legitimate business question underneath it. Someone has to choose the picture. Hamedi is asking that person to bring evidence to the choice. Beauty may belong to the beholder; the merchandising meeting would still like a number.

“Text kind of gets you found, but images get you chosen.”

Jehan Hamedi · Talk Commerce, October 2026

Capital follows the question

By October 2024, Vizit had announced a $25 million Series B led by Industry Ventures. Existing and new investors participated, and the company said it would expand its platform and enter additional markets. Customers named around the round included Mars, Heineken, and Ghirardelli. A question about images had become an enterprise software business with money committed to its expansion.

The reported customer outcomes were specific, too. Vizit cited a 30 percent conversion increase at Mars and a 10 percent rise on Amazon for Ghirardelli. Those figures describe company-reported cases, with their own contexts. They do help explain the commercial appeal: a brand already spending to attract shoppers has a reason to improve what those shoppers see on arrival.

In 2025, Hamedi joined AACSB's Influential Leaders class. His earlier recognition included the Boston Business Journal's 40 Under 40 and the Path to Purchase Institute's Ecommerce 40 Under 40. His work had crossed from marketing into computer vision, audience modeling, and patented software. The business school graduate had acquired a substantial second vocabulary.

An inventor in a business-school conversation

The patent record adds colleagues to a story that might otherwise sound like a solo pursuit. A March 2025 grant concerning multimodal machine learning names Hamedi alongside Elham Saraee and Zachary Halloran. Assigned to Vizit Labs, the invention concerns training and scoring across content representations. The dry language documents a team building machinery behind the sales pitch.

Today, he is founder and executive chairman and a named inventor on 47 U.S. patents. He also lectures on AI, content analytics, and ecommerce, and mentors founders as a MassChallenge judge. The settings vary, but they all require explaining how a technical possibility becomes a useful decision.

A Boston AI Week event scheduled for September 24, 2026, paired him with Marcel Rindisbacher for a university fireside chat. His continuing relationship with Boston University now includes being listed among its Alumni Shark Tank participants. There is a pleasing circularity here: the student who left with a business degree returns with a company and an argument about how people choose. Away from the technical vocabulary, a podcast biography supplies a more sociable interest: craft beer. Wisconsin has retained a representative in Boston.

The shopper still gets a vote

Hamedi's recent emphasis remains the first visual encounter. On LinkedIn, he has questioned the amount of attention devoted to getting traffic onto product pages compared with understanding what happens immediately afterward. That is the consistent concern beneath the changes in product language and technology: reaching a shopper is expensive, and consideration can be brief.

In a Talk Commerce conversation published in October 2026, he made his view of images succinct: “Text kind of gets you found, but images get you chosen.” He argued that an asset's ordinary file information says little about its ability to sell, and predicted that scoring product imagery before publication would become routine. That remains his forecast, rather than a completed industry transition.

He also resisted the idea that shopping agents will simply make all our purchases unsupervised. The human at the end of the process still matters to him. For someone whose career has moved through social listening, recommendations, and visual AI, that is a remarkably consistent subject of attention. The tools keep changing. The customer still has to want what is on offer.

Back in Wisconsin, the difficulty was getting a prospective buyer to notice a manufacturing business. Years later, Hamedi is asking a version of that question of a digital shelf. He has given it algorithms, audience models, and a company. Its final examination remains wonderfully ordinary: someone looks at a picture, then decides whether to keep looking.