Sou Hu Cai Jing·2026-01-27 10:45Core Viewpoint - Huatai Securities launched the "AI Zhangle 1.0" trading app, the first AI-native trading app in the securities industry, after 100 days of iteration, featuring three core task areas: Early Listening, Special Reminders, and Task Board [1][3] Group 1: Product Features - The traditional brokerage apps often have multi-level menus and tab structures, leading to high discovery costs and long response paths for investors. The AI Zhangle 1.0 version integrates frequently used functions into three core tools [3] - "Early Listening" organizes information and interprets directions, "Special Reminders" focuses on key market signals, and "Task Assistant" efficiently completes execution actions, aiming to shift from a "user finding functions" model to an "intention-driven" model [3] Group 2: Technological Advantages - AI Zhangle is not merely a simple call to a large model interface but is built on a collaborative system of a main agent coordinating multiple "expert agents." This allows for tailored responses based on user intent, such as stock selection needs [3] - The app aims to be a useful, user-friendly, and reliable AI investment assistant, focusing on enhancing decision-making efficiency and trading services [3] Group 3: Data and Compliance - In the investment field, while general large models are knowledgeable, they struggle with discerning true information from false, particularly in understanding the underlying logic of news and its market impact [5] - Huatai Securities emphasizes the importance of "high-quality data" and "industry know-how" in developing investment AI, leveraging its comprehensive data across the financial business lifecycle [5] - The AI Zhangle app incorporates extensive internal research reports and proprietary knowledge bases through post-training of large models, enabling it to understand investment logic and handle specialized queries [5] - The financial industry has strict security and compliance requirements, and Huatai Securities has implemented multiple firewalls in AI Zhangle, including sensitive word interception and real-time content auditing, ensuring authoritative and secure data sources [5]