Bflysoft Achieves Profit Last Year... "AI Training Big Data Success"

Bflysoft strengthened both its external growth and internal stability last year, focusing on big data for artificial intelligence (AI) training.

On the 31st, Bflysoft announced through a disclosure of a change of 30% or more in sales or profit structure that it recorded an operating profit of 260 million KRW on a separate basis last year. During the same period, sales reached 16.9 billion KRW, an increase of about 4% compared to the previous year. Net profit also recorded 770 million KRW.

The strong performance of Bflysoft this time was achieved along with the steady flow of the existing news data platform for public institutions, ‘iSurfer,’ and the rapid supply of the new platform ‘RDPLINE (Real-time Data PipeLine)’ to major domestic conglomerates.

Launched in October last year, RDPLINE is a news data and real-time data supply platform essential for generative AI training. It provides integrated access to about 700 million raw data secured through the largest media alliance in Korea, big data labeled and refined from this, and 200,000 articles generated daily.

Recently, as the global generative AI market has rapidly grown, the issue of copyright protection for training data has become a hot topic in the industry. RDPLINE not only enables avoidance of copyright issues but also solves the ‘hallucination’ problem, a common error in large-scale AI where false information is presented as fact.

Hallucination refers to an error phenomenon where false information is conveyed as if it were true. To reduce this, AI companies have recently been actively adopting ‘Retrieval-Augmented Generation (RAG)’ technology. RAG improves the accuracy of answers by referencing and reflecting external reliable data such as news from media outlets before generating responses.

A Bflysoft official said, “The consecutive supply of the RDPLINE platform to large corporations handling generative AI-based devices and services last year greatly influenced the improvement of our performance,” adding, “Since real-time news data is essential for avoiding copyright issues and product advancement, we expect continued interest from various companies in the future.”

This content was produced with the assistance of AI translation services.