"Finishing reading 10,000 research reports at night and still being able to iterate investment strategies during the daytime, representatives from E Fund, GF Fund, Fullgoal Fund, China Europe Fund, ICBC Credit Suisse Asset Management, Yongying Fund, Huatai-PineBridge Investments and other institutions have made statements, noting that AI is reshaping the investment research production line of the public fund industry."

From star fund managers to the "industrialized assembly line" of investment research, why are leading public offering funds flocking to "All in AI"?

"AI can process more than 10,000 research reports or announcement information every night, and iterate investment strategies during the day." At a recent sharing event held by China Asset Management, the relevant person in charge of the company said that the empowerment of AI has significantly improved the strategy update frequency of the multi-asset team. Taking the convertible bond strategy as an example, more than 4 iterations have been completed in the past year.

Technology-enabled investment research is not an isolated case among leading public offering fund companies. As the Action Plan for Promoting the High-Quality Development of Public Offering Funds clearly puts forward the acceleration of the construction of a "platform-based, integrated, multi-strategy" investment research system, public offering fund investment research is fully moving into the era of "industrialized assembly line" with the help of financial technology.

A survey conducted by reporters from China Fund News on nearly 20 leading fund companies shows that financial technology has been deeply integrated into the whole chain of public offering investment research. From data governance and factor mining to investment decision-making and transaction execution, and then to compliance risk control and performance assessment, intelligent investment research systems are being upgraded from back-office tools to front-office combat effectiveness. Some leading fund companies have even explicitly put forward the strategic direction of "All in AI".

From "Personal Experience" to "Organizational Assets", Strategic Upgrade Under the Empowerment of Financial Technology

The public offering fund industry is undergoing a historic transformation from the "star fund manager" model to the "platform-based investment research" model. In this process, the strategic positioning of investment research systems has been raised to an unprecedented level by all companies.

GF Fund positions the investment research system as the "strategic core" of the company's digital transformation. Investment research capability is the foundation for asset management institutions to survive, and the investment research system undertakes the key mission of converging this core capability from "personal experience" into "organizational assets". Specifically, three levels of strategic objectives need to be achieved: promoting data to be transformed into credible assets, precipitating investment experience into reusable methodology, and moving the investment research process from discrete individual behaviors to an orchestratable and traceable organizational closed loop.

China Merchants Fund has upgraded its technological capability positioning from "ensuring business operation" to an operational capability that "embeds in business decision-making, serves customer experience, and holds the bottom line of risk", and the investment research system is the core production platform for the company's digital transformation. Its construction has gone through the stages of informatization, technologization, and digitization, and is now moving towards a new stage of intelligent digitization.

E Fund positions the investment research system as the core strategic carrier supporting the platformization of investment research capabilities, adapts to the management mode of "large platform, small team", breaks down system data silos, and realizes the integrated closed loop of research, investment, transaction and assessment. Yongying Fund also regards the construction of investment research system as an important strategic content of digital transformation. Fullgoal Fund has established the transformation main line with the trinity of "data-algorithm-computing power" as the core, and positions the integrated intelligent investment research platform as the core productivity base for digital transformation and an important carrier for the industrialization of investment research.

In terms of start-up time, the earliest layout of leading fund companies can be traced back to the period between 2013 and 2019. Although the layout rhythm of each company is different, all of them take the investment research system as a breakthrough to promote digital transformation. In terms of investment, all companies have shown a continuous growth trend. Many companies including Yongying Fund, Huatai-PineBridge Fund, Fullgoal Fund, ICBC Credit Suisse Asset Management, Dacheng Fund said that the investment in investment research systems has increased year by year, and the breadth and depth of functions have been continuously strengthened. Taking Huatai-PineBridge Fund as an example, the total investment in investment research systems in the past three years has maintained an average annual growth rate of more than 15%, and the investment structure has shifted from "focusing on hardware + data procurement" to the trinity of "data + computing power + manpower".

From Research, Investment, Transaction to Assessment Management, the "Industrialized Assembly Line" of Public Offering Investment Research Takes Shape

Whether the investment research system can truly empower the business lies in whether it can be deeply integrated into every link of research, investment, transaction and assessment management. According to the survey, leading fund companies have made substantive breakthroughs at multiple levels.

First of all, the empowerment at the research end is mainly reflected in "reducing burden and increasing efficiency". China Merchants Fund has implemented a number of AI applications, with an average daily output of about 30 interview minutes and financial report comments covering 31 industries, helping researchers devote more time to in-depth research. The newly developed investment research Agent of China Asset Management has integrated more than 100 Skills (professional skill modules), covering multiple scenarios such as data crawling, industry analysis, financial modeling, and report writing, which improves the efficiency of researchers' information retrieval by more than 60%; the "five-element model" that its equity team has used for a long time has also achieved intelligent upgrading, most of the form filling work can be completed automatically by the system, and the compilation efficiency is improved by about 50%.

E Fund introduced that it launched the enterprise-level artificial intelligence infrastructure platform EWork in 2026 to promote the large-scale application of AI capabilities in all scenarios of investment research. Guotai Fund has built a three-level progressive intelligent investment research architecture of "data base - knowledge base - Agent capability layer". The bottom layer is a structured data system covering core dimensions such as A-shares and funds, and the middle layer is an unstructured knowledge base including research reports, roadshow minutes, etc., which supports semantic-level retrieval.

The Tianji Equity Research System of Yongying Fund covers research data collection and research content management from macro, industry to company dimensions, and covers core functions such as morning meetings, simulated portfolios, and individual stock recommendations, realizing closed-loop management of research work. Its Star-Pulling Fixed Income Investment Research System uses big data technology and intelligent rating model to realize the integration of rating and pricing, risk monitoring and strategy management through a microservice architecture. In the field of asset allocation, its Vientiane Multi-Asset Intelligent System operates in coordination with the Jingwei Investment Management System and the Mingjing Risk Management System, and the five major systems form an integrated investment research support system covering multiple businesses such as fixed income, equity, and asset allocation.

Harvest Fund has been laying out the digitalization of investment research since 2018, and in the past two years, it has further deeply implemented AI capabilities into investment research business. Harvest Fund believes that the application of AI in the field of investment research can be divided into three stages: production, interaction and decision-making. At present, it has exerted great value in the two links of production and interaction. At the production end, AI greatly improves the efficiency of information acquisition, processing and structured processing; at the interaction end, it realizes personalized and accurate push of information, and fund managers can quickly browse information highly related to their own portfolios.

Penghua Fund said that it focuses on the construction of digital basic capabilities, strives to improve the technical system supporting business operation, breaks down the barriers of information flow, and improves the efficiency and stability of overall operation. The self-developed intelligent investment research platform of the company integrates functional modules such as massive underlying data, public opinion monitoring, factor analysis, and portfolio risk management, providing real-time and multi-dimensional decision support for fund managers, and greatly improving the research breadth and depth of researchers.

Secondly, at the investment end, financial technology focuses on expanding the breadth and depth of decision-making, and has evolved from a single-point tool to a multi-strategy portfolio management system.

GF Fund has built a multi-strategy portfolio management system in the active equity field, realizing full-link data aggregation and modular layout through the fund manager's workbench, and solving the problems of income traceability and portfolio perspective through position allocation management and position holding management respectively.

In the field of fixed income business, GF Fund has achieved a high degree of online and intelligent process in the over-the-counter market and inter-bank business processes. The integrated deposit system upgrades a variety of scattered offline operations to an integrated full life cycle management mode. The new bond system opens up the whole process of the primary market, and the intelligent trading robot effectively improves the efficiency of repo and spot bond transactions.

Qu Jing, Chairman of the Quantitative Investment Decision Committee of China Asset Management, introduced that large language models have significantly improved the efficiency of code writing and IT development. The deeper change comes from the expansion of data scope. Large language models can read announcements, news, research minutes and industrial materials, extract new stock selection features from the corpus, which has low correlation with traditional price and volume strategies, and helps to improve the stability of the strategy base.

Fullgoal Fund has realized full independent R&D and application of special systems in the three core business lines of equity, fixed income and quantification, and gradually promoted the construction of an investment research system with the business form of "multi-product, multi-strategy, multi-manager collaborative system". China Universal Fund has built portfolio management for money market funds and bond funds, as well as operation support management for bond ETFs, realizing functions such as real-time valuation of money market funds, real-time position update of fund products, position structure distribution, and penetration of asset categories.

Invesco Great Wall Fund builds an investment research platform based on cutting-edge technologies such as AI and cloud native, taking front-end integration and interface integration as the core starting point to support multi-strategy businesses such as fixed income+, fixed income, ETF, pension FOF, active management and quantification.

"The investment research system is not a simple informatization back office, but a core production infrastructure that connects 'data - research - model - portfolio - risk - review'." The relevant person in charge of the Quantitative and Overseas Investment Department of Huatai-PineBridge Fund pointed out that its value lies on the one hand in standardizing the research process, making it reproducible and precipitable, and on the other hand in expanding the research coverage and shortening the research cycle through automation and intelligence.

The reform at the transaction end is also remarkable, which is mainly reflected in the great improvement of the automation level. Ping An Fund said that its investment research platform has gradually formed an integrated framework covering pre-investment research, in-investment support and post-investment analysis, and extended sub-systems such as investment management, research management, portfolio analysis and risk performance. Intelligent applications include intelligent instruction recognition, intelligent meeting, information summary and so on.

At the assessment management end, technology empowerment realizes traceability from qualitative to quantitative. Yongying Fund introduced that the management of research work is incorporated into the system, adding quantitative measurement standards for work, providing more reliable and objective indicators for team assessment management. Fullgoal Fund has established a relatively complete mechanism in performance attribution and investment research assessment management, providing systematic support in the breadth and depth assessment of researchers, performance attribution analysis and assessment of equity and fixed income investment managers.

Compliance and Risk Control Are Embedded in the Whole Process of Investment Decision-Making

Against the background of stricter supervision and the penetration supervision becoming the norm, technology has realized the pre-positioning of compliance and risk control and embedded it into the whole process of investment decision-making.

According to the introduction of China Merchants Fund, in its investment research system, it adheres to the principle that compliance and risk control rules take precedence over model judgment. For rules such as shareholding concentration, liquidity constraints, fair trading, investment scope and authorization quota, a professional rule engine is used to implement rigid control, and AI cannot modify, bypass or replace them on its own. The company has built multi-layer defense lines of pre-warning, in-process monitoring and post-event analysis, launched more than 60 self-developed risk control dimension pools, and the computing performance has been improved by about 30% after the cluster transformation of the real-time engine.

GF Fund takes compliance and risk control as a module that must be "embedded" at the beginning of system design, which runs through the whole process of pre-event, in-event and post-event. Risk control rules such as rating, concentration and qualification are automatically checked in key business links, real-time warning or rigid interception is prompted, and post-event data is synchronized and traced, so that the rules can be configured, the process can be traced, and the results can be traced.

Fullgoal Fund has built a compliance and risk control engine based on distributed EDA architecture, promoting the digitalization of the whole process of pre-investment risk control calculation, intraday risk warning and post-event compliance supervision. The self-developed risk control platform provides functions such as visual view of investment policies and standardized risk control rule configuration, supports pre-examination of intended transactions and simulated position adjustment calculation, and realizes full coverage of various on-site and off-site transaction scenarios and investment varieties.

The self-developed investment research system, risk control system and transaction system of Yongying Fund are all connected with the Hengsheng O32 system in real time, realizing the whole process management and control of investment research transformation, investment decision-making, risk control constraint and transaction execution.

In terms of investment concentration management, the investment research system gives the fundamental view and internal grading evaluation of the entity and individual bonds, the risk control system controls the investment quota according to the positioning of different product strategies, and the transaction system realizes pre-event and in-process control. In terms of portfolio liquidity management, the self-developed risk control system monitors the liquidity of each portfolio daily, and automatically alarms for situations such as too high proportion of restricted assets and mismatch of assets and liabilities. In terms of fair transaction management, the self-developed transaction system and O32 system embed standardized transaction entrustment and allocation rules, and the risk control system regularly carries out multi-dimensional fair transaction analysis.

Dacheng Fund adopts the pre-investment risk control method to pre-position the compliance and risk control constraints, and implements the management and control requirements through the full closed-loop management mode of pre-investment compliance verification, in-investment real-time interception and post-investment review and traceability. Huatai-PineBridge Fund divides risk control rules into "hard constraints" and "prompt constraints". It automatically restricts clear rules such as shareholding concentration, liquidity and investment scope when the portfolio is generated, and handles occasional events such as abnormal fluctuations through early warning and manual review.

ICBC Credit Suisse Asset Management promotes the full connection between the intelligent investment research platform and the intelligent risk control platform, pre-positions the compliance requirements to the links of investment calculation and portfolio construction, and realizes "compliance first, then investment". At the same time, it builds the IBOR real-time transaction data center to provide data base support for the real-time and rigid implementation of risk control rules.

In terms of data security, many companies adopt the principle of "privatized deployment first" to ensure that sensitive business data runs in an internal controlled environment. Hirano, Chief Director of Artificial Intelligence at J.P. Morgan Asset Management China, said that core models and sensitive data run in an internal secure environment to ensure that customer data and investment strategies are not leaked.

The Future Is Optimistic About the Integration of AI Agents and the Whole Process of Investment Research

Looking forward to the next three to five years, all companies have formed a relatively consistent judgment on the biggest technological breakthrough point of the investment research system: AI Agents will realize deep integration with the whole process of investment research.

"The biggest breakthrough will come from the integration of multi-modal AI and intelligent decision-making. The ultimate goal is to build a new paradigm of investment research with deep human-machine collaboration. AI is responsible for breadth and speed, and humans are responsible for depth and judgment." Hirano said.

Dacheng Fund believes that the biggest technological breakthrough point is the enterprise-level financial agent, and relevant technology and business breakthroughs should be carried out with the native integrated architecture of financial agents with full-link governability, collaboration and iteration. In the future, the agent architecture will turn the whole process of investment research into a complete digital production line with autonomous collaboration, human-machine division of labor, precipitation of investment research knowledge assets and embedded compliance.

Huatai-PineBridge Fund believes that future AI assistants will have deep personalized capabilities, be able to learn the investment style and preferences of fund managers, and provide customized decision support. Guotai Fund predicts that the differentiation will come from whether the organization can scale up knowledge, and precipitate the methodology, industry framework and compliance red line of senior researchers into standardized, reusable and evaluable Skills.

From the front-line practice at the investment end, the digital intelligent transformation of investment research systems is profoundly changing the working methods of fund managers and researchers. "High-quality investment results do not come from massive information, but from the right information, the right time and the right judgment, which is the core value of professional investors, and AI is the key tool to amplify this core value." Hirano emphasized.

ICBC Credit Suisse Asset Management said that the company has established an AI Application Engineering Department, formed a full-time FDE front-end agile combat innovation intelligent investment research flexible team to deepen the integration of business and technology. Bosera Asset Management believes that artificial intelligence will promote a substantial improvement in the efficiency of investment research, from quantitative change to qualitative change, affecting organizational reform and industry ecology reshaping. The company also collaborates all departments of the front, middle and back offices