From the experimental field to the main battlefield, brokerage firms are ramping up their in AI agents.

Since the beginning of this year, AI Agents have been rapidly penetrating the core business scenarios of securities firms. According to data from Analysys Qianfan, as of the end of July 2026, 32 securities firms have completed the filing of algorithms and models, with a total of more than 50 filing projects. Among them, over 10 leading securities firms have completed multi-model filings. Currently, AI Agents have been implemented in various aspects such as investment research, investment advisory, operations, and auditing within securities firms, with the application scope continuously widening. Experts interviewed believe that 2026 is a critical juncture for AI Agents in securities firms to move from pilot programs to large-scale validation. At present, the reasoning and tool-calling abilities of large models have significantly improved, enabling Agents to complete multi-step and relatively complex task chains; various securities firms are shifting from testing individual business points to a company-wide strategic layout, with substantial increases in investment in computing power, data, and talent, presenting extremely broad development prospects.