On June 16, Ant Technology officially launched the Agentar Financial Intelligent Agent Expert Group at the main forum of the China International Finance Exhibition. Targeting financial institutions such as banks, securities companies, and insurance companies, it covers core business scenarios including wealth management, financial risk control, and financial marketing, driving the evolution of financial AI from tool-based applications to position-level intelligent agents.

According to the introduction, the Agentar Financial Intelligent Agent Expert Group includes ten financial digital experts and over 300 industry intelligent agents. Each "digital expert" corresponds to a complete financial job role, no longer limited to single-task execution, but rather capable of understanding business objectives, breaking down complex tasks, coordinating across domains, and scheduling multiple AI assistants to collaboratively complete end-to-end business processes, thereby reducing manual intervention and improving overall operational efficiency.

image.png

Different from traditional AI tools and Copilot models, most financial AI is still at the stage of task assistance. However, the Agentar expert group further upgrades intelligent agents into "position execution entities." For example, in marketing scenarios, the digital customer group management expert can directly receive target instructions, automatically connect with multiple sub-intelligent agents for data analysis, market assessment, strategy formulation, and channel outreach, parallelly executing tasks and compressing the previously days-long marketing strategy process to be generated within minutes.

The core capabilities supporting this system include the task management mechanism and experience accumulation mechanism. The former gives each financial digital expert the ability to autonomously decompose tasks and dynamically schedule professional AI assistants, achieving end-to-end process orchestration; the latter transforms effective decision-making paths from real business into reusable institutional knowledge assets through long-term memory and skills accumulation mechanisms, allowing the intelligent agent's capabilities to continuously evolve with usage.

image.png

In terms of coverage, the Agentar expert group focuses on high-barrier positions in the financial industry, including investment research and advisory, risk management, anti-fraud, claims processing, and client managers. These positions generally rely on complex judgment abilities and cross-system data integration capabilities, requiring high professionalism and compliance from AI. For instance, in risk control and anti-fraud scenarios, the intelligent agent must complete non-standard identification and cross-system analysis within a very low tolerance space; in investment research scenarios, it must provide explainable decision-making references amidst conflicting multi-source information.

Currently, this system has completed full-process verification at a major domestic commercial bank. Data shows that after AI takes over many execution tasks such as data organization and cross-system queries in the client manager work scenario, the end-to-end processing efficiency has increased dozens of times, and the scale of client management has increased more than tenfold.