After securities firms rush to integrate DeepSeek and WorkBuddy, what will set AI offerings apart?

Cailian Press, October 4 (Reporter Chen Junlan) — Over the past two months, the most intensive move among securities firms has been "onboarding."

On September 2, Tencent's WorkBuddy open platform officially went live, and GF Securities became the first securities firm to join the ecosystem zone, with the two parties jointly launching the "GF Securities" Buddy application zone, debuting 12 self-developed Skills and 9 expert capabilities; on September 7, Alibaba's Qwen open platform launched more than ten financial agents, four of which—Guotai Haitong's "Lingxi," Industrial Securities' intelligent investment assistant, Soochow's "Xiucai," and CICC Wealth—were securities firm agents; in late September, the six major skills of CITIC Construction Investment's self-developed "Dragonfly Skill" also went live on the WorkBuddy skill plaza.

From integrating DeepSeek, to launching Skills, to securities firms recently joining WorkBuddy in droves, the securities industry's embrace of external large models and the Agent ecosystem is still accelerating. But a new question has emerged: as more and more securities firms use the same or similar general-purpose large models, what will ultimately set securities firms' AI apart?

It is not a relationship of substitution; the differences at three levels are what matter

The first is the knowledge base. The corpus of general-purpose large models comes mainly from the public internet, whereas securities firms' AI knowledge is built on domain-specific data such as market quotes, trading, holdings, investment research and regulatory standards, layered with the research frameworks and business experience that institutions have accumulated over time. For the same question, a general-purpose model mostly delivers a "summary of public information," while securities firms' AI can be connected to capital flows, industry chain positioning and historically similar market structures.

The second is the deliverable. General-purpose large models mostly exist as standalone Q&A, delivering information and content; securities firms' AI must be embedded in real business processes, operating together with permission systems, risk control rules, compliance review and accountability chains, delivering business results that are executable, traceable and accountable. This also explains why securities firms' AI ultimately tends to point toward order placement, market monitoring and post-market review, rather than stopping at a chat box.

The third is information sources and boundaries. The professionalism of the financial industry determines that compliance is the bottom line of innovation, and "reliable sources and controllable output" are the two gates that vertical models cannot bypass.

Sample one: GF Securities puts AI into a mature app rather than starting from scratch

In July 2026, GF Securities released the AI-native version of its E-Taojin app, "AI Lens," built on its self-developed "Tianji Zhirong" large model service platform, integrating RAG retrieval-augmented generation and Agent orchestration technology to form a closed loop of "planning—tools—memory—evolution"; on the enterprise side, it built the Skill Hub platform, accumulating 30 categories of standardized agent plugins, and laid out an MCP open workbench, launching more than 110 MCP services including account management, market information and data tools. What is distinctive about GF Securities is that it did not make AI into an entry point, but into a network.

On the client side, it puts AI into a mature app. In July 2026, GF released the AI-native version of E-Taojin, "AI Lens," centered on four types of capabilities—"focus, magnify, see through, look afar"—corresponding to the four high-frequency scenarios of stock selection, market monitoring, trading and Q&A. The trading scenario can recognize users' vague expressions and, after confirmation, assist in completing orders, moving AI from "assisted inquiry" to "assisted execution." Unlike some peers that build separate AI apps, GF chose to "upgrade in place," so tens of millions of existing users need not migrate.

English text is being prepared.

The content chain is a link that outsiders underestimate. According to the company's introduction to the reporter, GF Securities uses the "Jinchuangyi" all-media content operations center to connect the hot spot radar, intelligent creation, intelligent review, distribution center and data cockpit; on the E-Taojin app side, it has formed an all-day companionship matrix in which "'Taojin Morning Express' distills clues before the market, interprets anomalies and rotations during the session, and generates personalized post-market reviews based on watchlist holdings."

Sample two: CICC Wealth bets on buy-side advisory with a vertical large model

If GF focuses on path selection, CICC Wealth focuses on redoing services. According to the company's introduction to the reporter, its financial vertical large model is not a simple Q&A robot, but distills the mature methodologies of CICC's 300 senior researchers and nearly 3,000 frontline investment advisors, built on internal investment research data and real-time market quotes, aiming to solve the industry problems of "fragmented output" and "point-based distribution across multiple scenarios" in traditional AI services.

On the product side, the intelligent assistant "Xiao Jinn" focuses on moving from "Q&A" to "proactive service": it pushes a dedicated AI morning report in the morning, tracks individual stock anomalies during the session and proactively alerts, and when hot events occur, sorts out the complete timeline from policy introduction and industry reaction to capital flows; the "AI Zhitou 50" series, centered on clients' trading flows, covers algorithms, stock diagnosis, market timing, stock selection and holdings diagnosis. CICC Wealth said the company will soon launch version V13.0 of its AI APP, with the main theme of "relevant to me, proactive service," focusing on building an online buy-side advisory service system.

The ammunition comes from the business itself. As of June 2026, CICC Wealth's buy-side advisory assets under management had exceeded RMB 170 billion, and its product assets under management had exceeded RMB 530 billion. On the external ecosystem side, the CICC Wealth agent also appeared on the Qianwen open platform in September.

In early 2025, securities firms' focus was still on the speed of integrating foundation models such as DeepSeek and on local deployment capabilities. At the time, Huafu, Guojin, Guoyuan, Industrial Securities, Everbright, and Guotai Haitong, among other securities firms, successively disclosed that they had completed DeepSeek deployment, and GF Securities, CICC Wealth and others quickly followed; at that time, the main applications were still concentrated in areas such as knowledge Q&A, investment research, customer service, and R&D assistance.

By this year, competition began to enter a second stage. With the rise of Skills and Agents, securities firms began exporting financial data, research methods, and business tools to more AI entry points; the emergence of WorkBuddy further extended the competition into the external ecosystem.

But judging from the latest disclosed practices of GF Securities and CICC Wealth, "accessing external models" and "building internal vertical capabilities" are in fact not two mutually substitutable paths.

External general-purpose large models solve basic capability issues such as language understanding, reasoning, and agent orchestration; securities firms' internal models and financial capability layers are responsible for addressing data credibility, professional depth, business execution, and compliance boundaries. The former determines whether AI is "smart enough"; the latter increasingly determines whether it "can truly be used in securities business."

Especially in financial services scenarios, compliance remains a threshold that cannot be bypassed. CICC Wealth said its AI output is based on internal investment research data and reliable information sources, and controls model output through financial security mechanisms; GF Securities emphasized that securities firms' AI needs to operate together with permissions, risk control, compliance review, and accountability systems.

From this perspective, the securities firms' AI race is forming an increasingly obvious "two-tier structure": the bottom layer is the general large-model and Agent ecosystem, which is becoming more open and standardized; the upper layer consists of the data assets, research capabilities, business processes, client systems, and compliance capabilities that each securities firm has built over many years. As foundation models gradually become an industry "standard," what may truly be difficult to replicate quickly is precisely the latter.

Collapse; An unprecedented layout! Securities firms' international business capital increases are heading toward the RMB 50 billion scale, and overseas expansion is moving toward comprehensive services