Neurofusion adds investment research AI agent to Valley AI

| NBNTV = Joo Hyeon-ung |

Neurofusion, a financial AI platform company, announced on the 15th that it has integrated AI agent functions that support investment analysis and financial reasoning into its specialized financial platform, 'Valley AI.'

Valley AI is a platform that supports investment analysis based on refined financial data, such as financial statements, market prices, valuations, corporate disclosures, and economic indicators. Neurofusion developed the new features with a focus on increasing efficiency across the entire investment research process, including data exploration, collection, market monitoring, and management of analysis results.

The newly unveiled 'Investment Research AI Agent' consists of three core functions: Chat, Workflow, and Research Desk. Following private beta testing, the company first introduced Chat, an interactive research feature, and subsequently released Workflow and Research Desk in sequence.

Chat understands the context of questions based on the screen and data the user is viewing in Valley AI to provide analysis results. It can be utilized for corporate investigations, financial statement analysis, Discounted Cash Flow (DCF), multiple valuation, and market briefings, and is designed to indicate sources for key figures and claims.

Workflow is a feature that monitors the market according to schedules or conditions pre-set by the user. When specific conditions—such as a certain time, stock price fluctuations, or new disclosures—are met, it analyzes the relevant data and provides it via reports and notifications.

Research Desk is a feature that visualizes analysis results through charts, indicators, and commentary. It is configured to allow users to manage the judgment criteria and context used in previous analyses while updating content in synchronization with real-time data.

According to Neurofusion, the AI agent incorporates more than 530 financial data pipelines built by the company over approximately four years, 107 MCPs specialized for the financial sector, and 30 types of professional analysis skills reviewed by internal financial experts.

The performance figures were presented based on evaluation results configured by the company itself. Neurofusion explained that in an evaluation conducted last July using 84 financial-related questions, Valley AI's financial query analysis accuracy was 95.2%, while the compared general-purpose AI model recorded 78.6%.

In comparisons of high-difficulty financial tasks, Valley AI showed a win rate of 78%, and it added that response processing time and token usage decreased by 59.6% and 25.7%, respectively, compared to the comparison model. However, these results are figures calculated based on the company's own evaluation environment, questions, and comparison models.

Neurofusion plans to build a marketplace in the future where users can share and reuse queries, workflows, and research desks they have created with other users.

"Our goal is to enable more users to utilize refined financial data and analysis functions," said Choi Han-cheol, CEO of Neurofusion. "We will advance our features so that investors can increase research productivity while maintaining control over their final decisions."\

Meanwhile, Neurofusion was founded by CEO Choi Han-cheol, who has experience as a data scientist at global financial firms and a hedge fund trader. Valley AI, which officially launched in January 2026, has secured more than 10,000 paid subscribers and 30,000 free members, with users creating approximately 6,000 pieces of content every month.

The analysis results provided by the AI are reference materials to support investment decisions and are not information that guarantees profits or recommends the trading of specific financial products.