The launch and the business model shift
Wiseitec, a Korean company specializing in artificial intelligence and data and led by Kim Dasan and Lee Jaedong, presented its AI-powered YouTube analysis service called Tumeet on October 27 in Seoul during a press conference dedicated to the launch. ETNews reported it. According to the source, the product stems from the idea of applying to content planning and YouTube channel management the expertise the company has built in managing, analyzing and leveraging data in industrial contexts.
The strategic point is the transition from a business based on projects and custom implementations to a subscription service, with the stated aim of broadening the customer base to creators and businesses. The company justifies the choice with growing demand for more rigorous, data-driven channel management as YouTube consolidates its role as a central channel for corporate branding and marketing.
The stated features
According to the material, Tumeet collects and analyzes channel content and data daily to offer, across 13 categories, a ranking of the 100 most observed terms with related information on videos and channels. The company claims the service helps interpret the trend of interest by keyword, related terms and associated channels, because YouTube's personalized recommendation makes it difficult to grasp the overall market picture on one's own.
The service also includes custom dashboards, an AI-based operational support tool called Tumeet PD, and features for channel monitoring and reporting automation. These are capabilities described by the company and presented at launch, not independently verified results.
The cited case and its limitations
Park Shin-woo, team leader at Quan Entertainment, which manages Haha PD's YouTube channel, spoke at the presentation. According to the data shown on stage, the channel has 1.17 million subscribers and lost 68,953 subscribers in 475 days; analysis of entry and churn paths reportedly made it possible to distinguish content that brings in new viewers from content that encourages returns. Park said the choices were made with data, not intuition.
The most striking figures concern the JMM Expedition series: after an initial episode stuck at 225,000 views, the next episode reportedly exceeded 1.96 million in nine days, about 8.7 times as many; the series reportedly then totaled 12.02 million cumulative views across ten episodes, with 147,000 likes and 28,000 comments, and a 72 percent increase in average viewing time. Three Shorts derived from moments identified by the AI reportedly reached 15.82 million total views.
These figures were presented by the company and the client at launch, without comparison with control groups: on their own, they do not demonstrate a causal relationship between the use of Tumeet and growth. Prices, availability and the timing of the announced expansion toward businesses, institutions, MCNs and agencies also remain to be verified. Kim Dasan said the goal is to make data a foundation for editorial and channel-management decisions.
