Interview with Triplet CEO Donghwa Shin | "Know-how Must Be Digitized and Utilized"

Triplet is an AI startup specializing in AI-based video analysis and big data analytics. The company has deep expertise particularly in offline data.

It was founded by developing 'Papaboo,' a designated driver app that offers optimal pricing based on location data and conditions such as weather and day of the week. Since then, the company has expanded its scope and now provides solutions that handle a wide range of offline spatial data, including spatial safety data and retail data.

Through its AI-based retail solution platform 'DeepLounge,' the company has secured prominent clients in the distribution industry, including GS25, E-Mart, Lotte, and GUESS. It has also broadened its reach to various organizations and businesses operating offline spaces, such as Hallasan National Park, municipal libraries, wine shops, and women's fashion boutiques.

Earlier this year, the company participated in CES for the first time, and has recently appointed Samjong KPMG as the lead manager to recruit investors. The target investment amount is approximately 4-5 billion KRW.

The era of running businesses on intuition and gut feeling is over.

Triplet CEO Donghwa Shin began the conversation by describing the changing market landscape.

"For several years now, there have been constant warnings about the crisis facing offline stores. To make matters worse, the COVID-19 pandemic accelerated the 'retail apocalypse' phenomenon. While the pandemic may be ending, the retail apocalypse is still very real. Therefore, offline stores of all sizes must—and will—provide value and experiences that surpass what online can offer."

Indeed, movements reflecting this sense of urgency are gradually emerging. CEO Shin explains that an increasing variety of businesses across different industries and sizes are utilizing offline spatial data. He cited Triplet's municipal library and national park clients as examples.

"While data utilization was previously centered around large corporations, even small-scale stores are now leveraging data. RetroMoon, a women's vintage fashion boutique among our clients, is a prime example. RetroMoon operates not only an offline store and website but is also active on social media, selling products through live commerce streams. The store has such strong loyal customers that products regularly sell out during live sessions. Accordingly, RetroMoon aimed to transform its offline store into a space that provides brand experiences for loyal customers and delivers hyper-personalized services based on data. They are building services through kiosks that link website member data and recommend matching products. As times change and generations evolve, I believe more and more offline businesses will actively use data like this."

However, utilizing offline data is easier said than done. There are significant real-world barriers to entry. The initial deployment cost is likely the first concern, followed by the inevitable question: 'Does analyzing offline data actually increase sales?' This was exactly what Triplet deliberated most while building and improving DeepLounge.

"For offline businesses of all sizes to use data without barriers, cost and utilization challenges must be addressed. Triplet's DeepLounge connects multiple 360-degree cameras to cover wide spaces and flexibly uses edge computing or cloud solutions to reduce deployment costs. It also provides detailed data insights considering multiple factors—such as whether today's visitor count is above or below average, and how new product launches or marketing campaigns are performing. DeepLounge 2.0, with significantly enhanced reporting capabilities, is also coming soon."

"We are also focusing on hyper-personalization solutions that can generate immediate results. When recommending products and services tailored to individual customers based on offline store data, we can effectively increase purchase conversion rates and contribute to revenue."

This capability was the foundation for the strong reception Triplet received at its first CES participation in Las Vegas in January. The company achieved results beyond expectations.

"After registering your face at the facial recognition booth, standing in front of the wine recommendation kiosk or fashion item recommendation kiosk would trigger a message saying 'Hello, here are recommendations tailored for you, Mr./Ms. [Name].' It was a showcase of how spatial analysis, facial recognition solutions, and hyper-personalized services connect together. The response from attendees was outstanding. We were also able to establish a foothold for overseas expansion. A local liquor shop with more than 25 locations in the U.S. requested a meeting during the exhibition and completed due diligence. We are currently coordinating the deployment of our store analysis solution and wine recommendation AI kiosk."

Digitize your know-how.

CEO Donghwa Shin once again emphasized that offline data is the foundation of business and that the technology is now in place, so there's no need to overthink it.

"Whether online or offline, everyone running a business is working with data. For example, to effectively sell, you need to understand customers' purchase items, tastes, and preferences to develop effective sales strategies and reduce inventory and logistics costs. In reality, offline data isn't far away—it's your know-how. The problem is that offline data exists as 'know-how' rather than as 'data,' and such know-how is difficult to accumulate. It requires considerable time and effort. I believe that digitizing human know-how can become the foundation for optimal performance."

CEO Shin revealed that through DeepLounge, the company has achieved meaningful results: a 31.5% improvement in purchase conversion rates, a 14.4% increase in visitor growth, and a 9.2-minute increase in visitor dwell time.

He then cited Triplet's wine recommendation kiosk as an example of successfully digitizing know-how. While some may question how a machine can make recommendations, CEO Shin explains that an AI with 200,000 wine data points provides recommendation services, allowing people to focus on managing beyond that.

"The AI model in Triplet's wine recommendation kiosk learned wine data along with what people say about each wine. For instance, it collected situational data such as: Wine A's primary consumers are Korean women in their 20s, and they typically drink Wine A at birthday parties, pajama parties, and Christmas parties."

"As a result, it doesn't simply recommend based on 'it's a French wine with low sweetness and high tannin.' Instead, it makes recommendations based on contextual data—your upcoming birthday party, what food you plan to eat, your preference for rich chocolate, and so on. Even people who know nothing about wine can easily find what they're looking for. In the future, we plan to expand the recommendation service to include traditional Korean liquors."

Finally, CEO Shin urged that the need for offline data must be recognized quickly and, above all, that the mindset for approaching data management is important.

"I want to say that having an owner's mindset is crucial. Though I may be digging my own grave here (laughs), I believe that if you don't just stop at deploying a solution but actively demand and brainstorm with your provider about how to collect and utilize more data, you'll achieve far better results. Also, data might not make sense at first. But after 3 months, 6 months, 12 months, you'll be able to identify clear trends. While there may be short-term gains, maintaining a long-term approach to data will yield much better results."

Source: CIO (https://www.cio.com/article/3505490)