Molecular Scanning Brings Instant Coffee Quality Data to Producers and Traders

In 2016, Alan Lai was traveling through Uganda for work when he encountered chia seed farmers in the field. They labored in punishing heat and heavy dust, yet took home only a small share of what their product fetched at retail.

Lai was also a buyer, and he struggled to judge the quality of the chia seeds he wanted to purchase quickly and dependably. Despite how much of commerce had gone digital, he saw no accessible, objective method for rapidly establishing product quality—a gap that, if closed, could help farmers command better and fairer prices.

“A 1kg bag of chia seeds cost me about £.50 GBP [roughly $0.67 USD] while the very same seeds being sold at grocery stores in London, where I was based at the time, were being sold for £6 GBP [about $8 USD] for 250g,” Lai says.

That experience led Lai to found ProfilePrint, a Singapore-based technology company offering an easy-to-use quality-assessment tool for food products. The tool is also intended to lift producer profits by making quality measurable—what Lai describes as “democratizing quality assessment of food ingredients.” Since the company started in 2017, its products and software have been used in more than 80 locations on six continents.

Coffee is among the products it analyzes. Traditionally, Lai says, coffee quality has been assessed either by trained Q Graders or with laboratory equipment. Lab methods can be highly accurate, but deploying them is often costly and slow. Smaller players—coffee cooperatives, individual farmers, emerging traders—frequently lack the on-site access to reliable quality data that larger buyers enjoy.

“Democratizing quality assessment, for us, means leveling that access,” Lai says. “The system is portable and easy to operate, making it possible for users without QC lab skills or cupping experience to generate quality insights that were previously available only to the larger companies.”

How It Works

Picture a coffee trader in Nairobi handling multiple lots from different cooperatives. Under the traditional approach, evaluating each lot means roasting samples, running cupping sessions and shipping physical samples to overseas buyers—a process that can consume days or weeks and pile up costs. With ProfilePrint’s Mini Beluga tool, that trader can instead scan each lot repeatedly, generating a detailed digital fingerprint of a given coffee.

The recently released Mini Beluga uses spectroscopy to capture the molecular signature of a coffee sample. Green beans placed inside the device are illuminated, and it measures how molecules absorb and reflect light at different wavelengths. The result is a distinctive spectral pattern—a “digital fingerprint” reflecting the sample’s full molecular composition—produced in seconds.

That fingerprint goes to a cloud-based AI platform, ProfilePrint Lite, where machine learning models analyze it and map it against key quality attributes including SCA scores, flavor profiles and moisture levels. The company’s in-house Specialty Arabica Green Bean AI model draws on a global database of more than 30,000 specialty coffee samples assessed by Q Graders worldwide.

The advantages for traders are considerable. They can produce quality reports, compare lots against approved reference samples, and share findings directly with buyers and partners, which supports faster and more transparent transactions. Verifying shipment quality before export can also lower the risk of rejections.

ProfilePrint also runs an enterprise platform for larger organizations, but has lately shifted its focus toward small and mid-sized enterprises, buyers, roasters and regional traders. Getting started with ProfilePrint Lite is simple, according to Lai: users register for a free online account, then can subscribe to the Mini Beluga analyzer through a leasing arrangement priced at $500 per month. The minimum commitment is one year, and the package includes 3,000 credits annually; one credit lets a user turn a single green coffee bean sample into a digital lot that can then be run through AI models to view recommendations and predictions.

“The work the ProfilePrint team has done to support real-world integration has been exceptional, and is creating tangible benefits for our business,” says Carl Sara, global head of technology innovation at farm-to-roaster coffee company Sucafina. “What started as a device that provides us with data has grown into a tool that supports decision-making and speeds up our team’s work, letting our people focus on what they do best: making the calls that really need that human touch.”

Expansion to Africa

Several chain players view the technology favorably. One is Lisper Ndung’u, chief executive officer of the Nairobi Coffee Exchange (NCE), who has welcomed it and said it is likely to cut coffee sampling costs substantially at the exchange’s central room in Nairobi.

Lai says the capability carries particular weight in African coffee-producing countries, where output is plentiful but access to quality-assessment tools is scarce. By making analysis objective and fast, the company helps producers better capture the true value of their beans and compete more effectively in global markets.

“In Africa, adoption is gaining momentum, particularly in key coffee-producing regions. We’ve already deployed our solution with exporters in Kenya, Tanzania, Uganda, and Rwanda,” Lai says. “Several traders, cooperatives, and exchanges have shown strong interest in using AI-driven quality insights to improve transparency and reduce transaction friction.”

To Lai, transparency means buyer and seller can see the same quality results for a given coffee sample—the same molecular fingerprint and the same predicted quality characteristics—“which removes a lot of the ‘he said, she said’ that can creep into traditional quality disputes.”

Lower transaction friction benefits buyers as much as producers. When buyers can assess quality before goods physically arrive—even before payment—they gain far more confidence in what they are committing to.

In March, ProfilePrint held a demonstration in Nairobi, a notable milestone in its push into Africa. Lai sees Kenya as a strategic entry point given its well-established auction system and strong quality culture.

“Looking ahead, we see strong potential across Africa’s major coffee-producing regions,” he adds. “With growing global demand for high-quality African coffees and rising price incentives, the adoption of accessible quality assessment tools is likely to accelerate.”

Overcoming Challenges

As with any frontier technology, earning trust is among ProfilePrint’s main hurdles, Lai notes. Coffee is a deeply sensory, experience-driven industry, so adoption depends on showing that technology can complement human expertise rather than replace it.

The most frequent misunderstanding, according to Lai, is that the company wants to dictate how people run their businesses, or that it believes its AI outperforms people with decades of experience.

“We’ve had people telling us straight, ‘I’ve been cupping for 20 years, and no machine is going to tell me what good coffee is,’” says Lai. “And they’re right! No sensor, no algorithm, and no AI model is going to replicate two decades of tasting experience, of understanding coffee and the market, of knowing what their customers want. However, what AI excels at is processing complexity quickly. We take a fingerprint of a sample and simply tell you, objectively and consistently, what’s in front of you. Not whether it is good or bad; that’s your call.”

ProfilePrint is not built to replace traditional cupping, Lai says. He maintains that human expertise stays central to coffee evaluation, especially when it comes to interpreting flavor and market preferences.

“When things fall short, it’s usually for one of two reasons,” Lai says. “The first is data: insufficient or inconsistent samples or data make it hard to build robust models, and good data is the backbone of any good AI. The second is a mismatch between human expectations and what AI can do, for instance, expecting a model to be 100% accurate, which isn’t realistic since humans aren’t 100% accurate either. Our role is clear: We provide the data, and the human experts make the call.”

Operating in regions with differing levels of digital infrastructure also poses practical difficulties. “We have addressed this by designing solutions that are portable, user-friendly, and accessible, even in resource-constrained environments,” Lai says. “Ultimately, our progress has been driven by close collaboration with industry partners and a strong focus on user needs.”

With deployments at more than 80 locations worldwide and across multiple food sectors—coffee, tea, cocoa, juices and others—ProfilePrint has recorded strong year-on-year growth. Lai says the company has seen 90% sales growth across key markets, with usage expanding exponentially from 2024 to 2025. Many of these deployments are now fully operational, having moved past testing or pilot phases into day-to-day decision-making.

Customer Reception

At World of Coffee Bangkok in May, three ProfilePrint users discussed their experience with the technology: Martin Sunghwa Jeong of Foosung Corporation, a green coffee importer and exporter based in South Korea; Fajar Maulana Fikri, director of coffee at Espresso Lab, which runs coffee shops and a roastery in Dubai; and Hemanta of the Nepalese Coffee Project.

“What defines ProfilePrint is speed, efficiency, and prediction,” says Jeong. “My favorite part of the Beluga and Mini Beluga is the matching function, because when you want to find some green coffees, it is really helpful for the importers and exporters, and also for the farmers.”

For Fikri, the technology has eased his workflow considerably, especially when Espresso Lab has many green coffee samples to cup. His team no longer needs to roast all of them, he says; they can scan, narrow the field, and move ahead with the samples that interest them most.

“The technology is fast, convenient, and in demand,” says Hemanta. “I love how the profiles are available online anywhere for you to check out.” Hemanta notes that Nepalese coffee farmers would likely benefit, since many have limited access to coffee labs or to tracking their coffee’s quality from one year to the next.

For Lai, such use cases illustrate the technology’s potential. “What AI offers is speed, consistency, and objectivity,” he says. “It enables users to process complex molecular data instantly, providing a reliable baseline for decision-making. By combining human judgment with AI-driven insights, the industry can achieve greater efficiency, reduce subjectivity, and improve trust across the value chain.”