Recently, the 2019 Guangdong-Hong Kong-Macao AI Enables Finance Business Development and Innovation Conference was held in Guangdong. The People’s Bank of China – Guangzhou branch, China Insurance Regulatory Commission – Guangdong Bureau, commercial banks from Guangdong-Hong Kong-Macao greater bay area, and other financial institutions, the representatives of IT enterprises, total more than 130 people attend the meeting to conduct in-depth discussions on the topic of “how AI technology can enable finance business to develop”, collaboratively exploring the development direction and opportunities of finance industry in the AI era.

The senior researcher from the WeBank AI department, Dr. Quan Li was invited to give a speech on the application of federal learning and AI visual analysis technology in the corporate financial scenario. Introducing the intelligence application of retail scenario, Quan Li shared WeBank’s recent experience of practice application in the technology integration and Fintech innovation.

 

The growth of the retail industry is limited, AI may be the key solution

In the last few years, subject to the high labor cost, the difficulty in transferring experience, inaccuracy of sales forecasting and other reasons, the retail industry is facing the increased challenges. The most obvious problem is that the human factor accounts for an excessive proportion of the whole retail operation process. As a random variable, the human factor is difficult to be predicted and controlled. It is a high labor cost for companies and the performance of employees cannot be guaranteed. Therefore, it is the great significance of the transformation from the labor-intensive retail industry to the technology-intensive retail industry.

One classical scenario is that to clean up 60 million sale data samples, sort out those samples, and then use the labor model to predict the popular products, the performance of 1000 persons may not be better than that of one people subject to the limits of one’s time and energy. However, it will be an overgeneralization if you make decisions by your experience because you only know and use a small portion of the data to make decisions on product selection. But AI can finish the jobs that cannot be completed by manpower. It also can exploit the data to drive company decisions on operation in a faster, wider, and more accurate way, helping the traditional industries to address these challenges.

 

WeBank AI develop smart retail to help digital upgrade of enterprises

In the meeting, WeBank introduced the technology of “AI + Vision”. The retail industry can lower the cost and improve the efficiency in the key part such as intelligence warehouse and site selection.

Taking the intelligence warehouse location selection of the retail industry, the traditional method of location selection relies too much on the manpower. The manual site selection is not only a heavy workload, a time-consuming and long cycle, but also highly dependent on personal experience and insufficient data support. However, when we conduct the site selection, we may meet the difficulties of data sample amount limits and inaccuracy data.

AI perspective data and visualization analysis is the function of intelligence site selection. It can analyze the distance of warehouses to the business district, high-speed intersection and the wholesale market. Conducting the evaluation automatically of the traffic, the environment quality to make the decision of intelligence warehouse site selection. In this way, the company efficiency of the traditional method – site selection is increased by 25%. Compared to the old method, the delivery cost and rent cost is declined by 13% and 7% respectively. Hence, the profit margin of enterprises and the market competitiveness is further improved.

However, intelligence site selection can help the companies improve the efficiency and lower the cost, which is the tip of an iceberg in the progress that AI help develop the finance business and innovation. The smart retail – shop OS platform presented by Dr. Quan Li, is based on the multiple technologies of federated learning, data visualization analysis, and blockchain. Dr. Quan Li proposed the smart retail solution contained by three core parts: intelligence operation, smart labor and inclusive finance, to meet the traditional retail business needs of strategic planning, supply chain management, marketing decisions, logistics operations, human source efficiency improvement and management and control of funds. This can enable AI to participate in every part of company operation to lower the proportion of manpower and realize the digital upgrade.

 

AI for goods: have both economic and social benefits

Integrating AI into smart retail production can reduce labor costs and improve production efficiency for enterprises. As for the employees, it can also improve effectiveness and create higher value. It is reported that only one application of smart labor can improve efficiency by 50% in the field of the retail industry and the utilization rate of effective working hours of nearly 80% positions is increased significantly.

Helping the production and creating higher effectiveness of the economy is the first step of AI commercialization implementation. After the full implementation, AI will assist humans to finish more and more works through the smart robot and automatic production and other methods. It can further free the manpower and create a huge benefit for society. As an important force in the field of the finance industry and AI industry, WeBank will cooperate will more academia and industry to promote the development of standard theory and practical application of federated learning technology. With a two-pronged approach, we can realize the great vision of “AI for goods” early.