Zhi DongXi (public number: zhidxcom)
Compile the | Zhao Di
Edit | Yunpeng
Zhidong February 11 news, according to foreign media reports, in recent years, Microsoft's profits, market value, etc. have seen a substantial increase, but the number of personnel in its financial team is basically unchanged, maintaining about 5,000 people, and its financial work to achieve digitalization and intelligence is the key to the company's stable number of financial personnel.
Recently, Cory Hrncirik, head of Microsoft's "Modern Finance" project, introduced to the Wall Street Journal how to improve financial efficiency and control the number of financial personnel through AI, cloud computing, machine learning and other technologies.
At present, Microsoft has made financial forecasting, risk management, and internal inquiry work digital and intelligent. In the future, Microsoft may focus on the application of AI, machine learning and other technologies in a more refined degree of work.
First, using machine learning to make financial predictions only takes 30 minutes, and the error rate is doubled
Seven or eight years ago, Microsoft began to think about how to automate financial work, improve financial efficiency, and enable employees to better focus on areas that cannot be solved with technology, such as negotiating with business partners, looking for greenfield investment opportunities, and so on.
Microsoft first introduced machine learning technology in financial forecasting. Financial forecasting refers to the work of financial workers to predict the future financial status of enterprises according to the recent financial activities of enterprises, using mathematical statistical methods, combined with current forms and future trends.
Before the introduction of machine learning, Microsoft's finance team often took 1,000 people three weeks to complete each quarterly financial forecast, but after the introduction of machine learning models in 2015, it basically took 30 minutes to complete quarterly forecasts and reduced the error rate from 3% to 1.5%.
Second, risk analysis, internal inquiries, Microsoft's many businesses to achieve intelligence
In addition, Microsoft has also realized intelligence in defining risks and digitizing risk management processes. According to Hern Sirik, Microsoft is currently using techniques such as machine learning to analyze documents from governments around the world to predict potential economic risks. Currently, the company is also working to use it for internal audit-related work, such as identifying which invoices can enter the automated process and which invoices require manual intervention.
▲Microsoft's "Modern Finance" official website
Before the introduction of AI, machine learning and other technologies, Microsoft finance personnel need to manually enter 50 data sets, and then analyze, and finally come to the results, now, financial personnel can store more than 100 data in the cloud, through cloud computing and other technologies automatically generate analysis reports.
In addition, Microsoft is also trying to introduce robots, setting up AI agents. AI agents use natural language processing related technologies and can understand more than 60 languages. It infers the other person's intention to speak and condenses the entire conversation into simple sentences. Currently, about 30% of Microsoft's internal query work is handled through AI agents.
Conclusion: Business intelligence improves work efficiency, and AI gradually becomes the key to enterprise digital transformation
In January this year, Microsoft released the second quarter of fiscal 2022 financial report, its revenue, net income and operating profit have achieved an increase of more than 20%, but in recent years, the number of Microsoft financial personnel has not increased significantly, this part of the efforts made by Microsoft in the digitization and intelligence of financial work.
At present, Microsoft has used AI, machine learning-related technologies to achieve efficiency improvement in financial forecasting, risk management and other aspects, but Herne Sirik said that in the deeper and more refined level of work, the performance of machine learning and other technologies is not very good, and the degree of digitalization and intelligence of financial work needs to be further improved to adapt to the wave of digital economy.
With the development of science and technology, AI plays an increasingly important role in the overall digital transformation of enterprises, in the future, how to better achieve industrialization of AI, in the process of digital transformation of enterprises, to empower the efficiency of more businesses, it is worth looking forward to.
来源：The Wall Street Journal