The trend of changing expectations of financial services along with the high speed of technological development make customers increasingly desire to have a better experience, with the products consistent with the needs, which does not require much human intervention. Knowing data resources and analysis technology, banks will have more opportunities to get closer to customers, thereby exploring new sources of revenue.
The banks are expected to be less of a direct contact with customers in the future, this comes with the need for management attention to be especially concerned, in the context of digital banking development. Not only applying and taking advantage of 4.0 technology achievements such as artificial intelligence (AI) and Big Data but, according to experts, in order to maximise the efficiency, the banks should take into account the harmonious story between risk management and technology application.
In the field of finance and banking, especially credit scoring, AI and Big Data are expected to assist the banks in reducing risks as well as saving maximum time processing information to give valuable signals during its operations.
However, according to the expert group from Tien Phong Commercial Joint Stock Bank (TPBank), it is not easy to build and maintain a team of people with in-depth knowledge of AI and Big Data, so banks should consider investment into the tools and solutions available to assist in automating the estimation, construction of credit scoring models on the basis that users only need to understand the principles and parameters of the model. It is not necessary to be an expert in data science to develop algorithms from the beginning, then to focus on building internal resources with professional capacity.
The way to get the data is also the problem that banks are copping with. The fact that customer records or credit history data are incomplete also make it difficult to build models with traditional methods, or for new products, the banks will not have enough or full credit history to assess risks.
In addition to the demographic data and history of traditional credit relations, TPBank experts found that some non-traditional information channels were gradually becoming popular in the field of data science that banks or financial companies can exploit to build models, including social network data; data on behavioural psychology; telecommunication data and mobile device usage data. These were the four data groups considered to be the best data sources currently in risk assessment, but the effectiveness of good or bad customer prediction of these data groups was uneven.
As such, the banks will have to select products and solutions with appropriate data sources from time to time. In addition, they can also take advantage of data from Fintech companies in terms of technical knowledge and model building data for the field of using technology to solve financial problems.
When it comes to risks, the prerequisite is to have appropriate policies. Pham Anh Tuan, the Board Member of Joint Stock Commercial Bank for Foreign Trade of Vietnam (Vietcombank), shared that it was necessary to have policy frameworks for banks to set up relevant implementation procedures throughout the process from the first time until the customers transact with the banks. Policies and procedures for identifying customers need to be flexible and support the fastest in providing services to customers while ensuring the factors of safety, security and anti-money laundering.
In addition, the process and policy should include an internal process (based on the governing body’s regulations), to identify and report suspicious bank transactions; control and compliance functions in accordance with international practices and standards.
Set in the process of digital transformation of banks, each bank itself must have a policy to classify the data appropriately to apply the new trend. According to a representative of Vietcombank, there is a need for internal processes related to data storage, data processing and data transmission. Storing data or a part of data on cloud also requires processes and regulations corresponding to the management and processing of data.
“The issuance of corridors and appropriate policies (corresponding to types of cloud) will help enterprises to conveniently deploy and apply cloud computing technologies,” Tuan said.
In general, new technologies and rapid changes in technology always require appropriate security policies. Security guidelines should include database and information security, eKYC security, and widespread cloud application. In parallel with that is taking advantage of security standards (ISO/ IEC27001: 2013, ISO/ IEC 27017, ISO/ IEC 27018) used in relation to cloud requirements for both providers and users. And the banks’ cooperation with Fintech companies, or the risk management when bringing products to market quickly must also have attached policies.
