Pint Applies 'Generative AI LLM' Technology to Robo-Advisor Engine

December & Company, which operates the AI simple investment platform ‘fint’, announced on the 13th that it plans to apply generative AI large language models (LLM, Large Language Model) to its representative stock investment algorithms for U.S. and Korean stock solutions. It is expected that the smarter AI will bring further improved performance, including additional returns and convenient usability.

Fint’s U.S. and Korean stock solutions are listed on the Koscom Robo-Advisor Testbed, respectively in the 2nd and 19th rounds, and have already achieved good results with long-term high returns. The cumulative return of the U.S. stock solution is 133.02%, while the Korean stock solution is recording 14.78%.

This is based on Fint’s proprietary AI technology accumulated over many years. Fint’s stock investment solutions incorporate a large amount of high-tech AI technology. Using Fint’s unique AI technology that applies ensemble methods to technical models based on stock price data and fundamental data models such as financial statements, it assesses the growth potential of individual stocks by utilizing advanced AI technology. Stocks with a high likelihood of showing good performance over the next six months are included in the portfolio, and real-time rebalancing is conducted. This method combines proprietary deep learning technology for large-scale data with factor analysis techniques to derive a stable solution.

Once the new LLM model is applied, it is expected to achieve more meaningful investment results. December’s LLM analyst stock scoring engine analyzes and calculates major trend changes in financial statement items and key financial ratios to predict whether future returns will increase. It features the implementation of the analyst’s inference process for forecasting company performance using Chain-of-Thought (CoT). By sequentially reproducing a human-like flow of thought, it is designed to predict whether a company’s stock will rise in the future. Users can also easily verify the AI engine’s inference process. Significant improvements in returns were confirmed in December & Company’s own backtest simulation results.

December & Company CEO Song In-sung said, “By adding generative AI LLM technology to the Isaac investment engine, which has already been delivering good results with AI technology that selects high-growth individual stocks, we expect to further enhance the performance of the investment engine.” He added, “We believe we can demonstrate new innovation and differentiated competitiveness in AI-based asset management services, which are gaining increasing attention recently.”

December & Company holds the No. 1 market share in Korea’s robo-advisor discretionary accounts with 80% of subscribers and 59% of assets under management. In addition to fint, it is expanding various partnerships and collaborations with securities firms and asset management companies, including asset allocation solutions, fund advisory, and direct indexing services. To provide robo-advisor retirement pension discretionary services scheduled to be implemented by the end of this year, it has secured partnerships with a total of 10 retirement pension operators, including three major large banks and seven major securities firms.