Application of digital intelligence in financial industry
How to integrate data intelligence and finance perfectly, improve the service efficiency of financial institutions, expand the breadth and depth of financial services by reducing costs and increasing efficiency, comprehensively empower financial institutions, and realize the implementation of intelligent, personalized and customized financial data intelligent applications of financial services, which are Zhejiang University System financial intelligence joint laboratory Gao Peng, pan Jing and others have been studying and pushing Entering a new era of intelligent finance.
In this report, Gao Peng, pan Jing and others will start with the essence of data intelligence and review the historical process of data intelligence. In view of various problems encountered by banks and financial institutions, they will try to provide quick win strategies and coping strategies under the current market competition pattern, as well as the application of data intelligence in the financial industry in the future, so as to help financial institutions better control data intelligence, Leading the data intelligence era.
u2589 what is data intelligence
From the perspective of technology, data intelligence refers to the scenario business flow driven by machine decision-making based on big data engine, integrating cloud computing, big data, knowledge map and large-scale machine learning, which can be called data intelligence. It can process, analyze and mine massive data through large-scale machine learning and deep learning technology, extract valuable information and knowledge contained in the data, make the data have intelligence, and seek solutions to existing problems and realize prediction by establishing models.
In short, data intelligence is the integration of big data and artificial intelligence technology. With the rapid growth of data volume, data intelligence will be directly applied to scenarios to improve operational efficiency. Its ultimate goal is to use a series of intelligent algorithms and information processing technology to achieve human in-depth insight analysis and summary under the condition of massive data, draw conclusions, and finally realize intelligent decision-making.
(Chart: Simulation of big data without photo disk system architecture)
(Chart: the development of data Intelligence)
The third stage: the combination of data and business scenarios, in the era of data mining and data modeling, big data companies started the construction of data platform, internalized AI modeling platform into their own capabilities, and formed solutions based on AI modeling platform to help enterprise customers implement big data application. At the same time, it has been greatly developed in this stage;
The fourth stage: big data driven machine automatic decision-making stage, so that the machine has reasoning ability, data intelligence embedded in the scene, directly driving business. With the gradual development of large-scale data processing, data mining, machine learning, human-computer interaction and NLP technology, this is the realization of advanced, real-time machine decision-making ability. The emergence of data intelligence will bring business and business model reconstruction.
In the future, with the technology becoming more mature, many execution links can be realized by machines. However, only when data intelligence has the ability to actively drive business can it be regarded as machine automatic decision-making, which means that data intelligence has entered a new level.
Finance, as the core of modern economy, gathers massive scene and user data. Meanwhile, the business has the characteristics of standardization. It is one of the industries most in line with the conditions of digital transformation. At present, almost all the technology giant companies are laying out finance, and the vast majority of banks are implementing the digital strategy.
There are three reasons for the financial industry to become the market for the giant technology companies to seize the layout: first, the financial industrys assets, transactions, products and even users are in the digital field. The definition is vertical and clear, and there are many raw materials that can be calculated. Second, finance is the most frictionless field, where there is no production, storage and logistics. Once the data closed loop is formed, the efficiency will be greatly improved; third, the financial industry is a trillion level market, and the runway is wide enough.
Multi technology first. At present, many financial big data companies have the ability of deep learning, NLP, knowledge map, which were originally considered as AI technology; from the perspective of customer demand, in order to guide decision-making, they need to gather massive multi-source data, which inevitably involves the processing of unstructured data. Based on complex network reasoning and decision-making, single technology is difficult to solve the problem;
Second, the formation of a new business model. When data intelligence can be embedded in the scene and support the automatic decision-making of machines, the operation mode of the whole financial industry will change greatly. From the original data aided human decision-making, it will be upgraded to the full-automatic and unconscious personalized business used by customers in the scene. Therefore, the industrial form and operation mode of financial institutions will change greatly, and even be spawned New species.
(Chart: new business model in the era of data Intelligence)
Among them, Xiangxiang big data provides three major engines, namely, real-time recommendation without photo disk, business opportunity mining engine without photo disk, and NLP without photo disk. Based on the closed-loop mode of brain + tool + scene, based on the big data basic platform, combined with marketing + RPA, and using cloud computing, big data, knowledge map and large-scale machine learning and other technologies, the collected customer transactions, consumption, network browsing, etc For data, it uses deep learning related algorithms to build models to help financial institutions connect with channels, personnel, products and customers, so as to cover more user groups and provide consumers with personalized and accurate marketing services.
u2589 implementation of data intelligence in financial field
In the financial industry, data can be used to predict, judge, make decisions, analyze, etc. the risk control, intelligent marketing and anti fraud of banks are all realized through digitization and AI. The implementation of data intelligence in the financial industry will bring about the following new changes:
1u3001 In the future, in the new finance, data will be bred into new production factors. Data has increasingly become the core factor of production and added to financial products. In the context of new finance, real-time recommendation based on scenarios enables people, products and scenarios to be fully and automatically scheduled. To a large extent, data intelligence will bring about the improvement of marginal efficiency and total factor productivity;
2u3001 In the new finance, data intelligence will bring greater multiplier effect. In the process of digital operation of financial products, the data is further directly driven to the scene, which greatly breaks through the spatial and temporal constraints of financial products and forms the birth of intelligent dynamic financial products;
3u3001 Data intelligence helps users to achieve multi product portfolio planning, so that everyone can have an AI financial advisor. AI can provide personalized services to everyone in real time, which can effectively break through the service limitations of financial long tail users. In conclusion, data intelligence can support the rapid development of digital finance, promote the product change, efficiency change and dynamic change of new finance, and provide new technical support for Chinas financial digital intelligence.
The no photo disk of big data in the financial industry is a fully automatic operation platform driven by data intelligence. It forms a full closed-loop process driven by data intelligence through four functions: automatic customer mining, automatic customer touch, automatic customer follow-up and automatic data backtracking. The whole process does not need human intervention. All resources are allocated and parameters are optimized based on data. The specific automatic process is as follows:
S1 automatic mining: the massive data of the bank is used for algorithm mining, and the push logic is extended to the event flow like timeline for mining. Finally, the deep learning and verification of users multi-dimensional characteristics is completed, and the data of different potential customers of each specific business are mined;
(Chart: simulate big data without phase disk S1)
S2 automatic access: for each potential customer group with different specific business, Di will correspond to different business recommendation strategies, build the product recommendation knowledge map of the whole business system without a photo disk, and use the way of telephone robot to reach and market users;
(Chart: simulate big data without phase disk S2)
(Chart: simulate big data without phase disk S4)
All the above four links are automatically driven by data intelligence. The four links can connect 100 million user contacts, penetrate into various channels such as bank outlets, telephone marketing, network marketing, wechat, etc., which can effectively improve the operational efficiency of banks by more than 30%.
u2589 new business model in the era of data intelligence
First of all, in the era of data intelligence, there will be a new business model.
As data intelligence is deeply embedded in business scenarios and processes, in addition to the original technology enabling mode which mainly provides solutions, in the era of data intelligence, a new mode of cooperative sharing and joint operation will gradually appear, which will greatly improve the ceiling of data intelligence companies. Cooperation sharing means that data intelligence companies can obtain the business budget of enterprises, not just it budget, which can greatly improve the ceiling of data intelligence companies in a single industry.
Cooperation sharing will be a new model. Data, technology and application scenarios are combined to give intelligence to the scene by machine decision-making, so as to improve the operation efficiency of the whole scene, so as to realize the cooperation sharing with enterprise customers.
Cooperation sharing improves customer stickiness, which is conducive to data intelligence companies based on the industry. Cooperative sharing means that data intelligence companies go deep into business scenarios. The ability of data intelligence companies to understand customer application scenarios is close to enterprise customers and far beyond other suppliers.
In order to implement the new business model of data intelligence, such as cooperative sharing and joint operation, three conditions should be met:
1. Only by doing incremental market can we cooperate and share. For enterprise customers, the value of profit center is greater than that of cost center. Therefore, to help enterprises develop new business and improve the capacity of the original business, we can carry out cooperative sharing. To help enterprises reduce costs is impossible to cooperate, because there is a clear ceiling.
2. It is necessary to build a full scene coverage covering cloud edge end. Only by covering the whole scene and implementing the full closed-loop of data can the value of data intelligence companies be proved, and the results can be quantified. Based on the quantifiable results, data intelligence companies can cooperate with enterprise customers to share.
In the financial field, data and algorithms are deeply bound into business scenarios to make the data itself a part of the business, such as real-time recommendation ability, risk control ability and dynamic interest rate calculation ability in the scenario, so as to realize the continuous improvement of business income and even spawn new financial business. Gao Peng: an expert in artificial intelligence, founder and chairman of big data of simulacrum. He studied under academician pan Yunhe. In 1998, he won the doctors degree of Zhejiang Universitys School of computer. He is the founder of the first set of mobile boss billing system model in China. Pan Jing: Director of foreign cooperation and Exchange Department of Zhejiang University, once led the Zhejiang university student team to participate in the student entrepreneurship (SRTP) project, and obtained the Universitys respectively Gold medal and other honors within the province source: public news editor: Chen Tiqiang_ NB6485
In the financial field, data and algorithms are deeply bound into business scenarios to make the data itself a part of the business, such as real-time recommendation ability, risk control ability and dynamic interest rate calculation ability in the scenario, so as to realize the continuous improvement of business income and even spawn new financial business.
Brief introduction to the author
Pan Jing: Director of foreign cooperation and Exchange Department of Zhejiang University, once led Zhejiang university student team to participate in the college student entrepreneurship (SRTP) project, and won the Gold Award in the University and the province respectively