Artificial intelligence will become the mainstream of financial services in the short term. In this survey, 85% of the respondents are using some forms of artificial intelligence, among which financial technology enterprises are slightly ahead of traditional financial institutions in adopting artificial intelligence. The trend of large-scale application of artificial intelligence already exists. Among the leaders of artificial intelligence, half have applied artificial intelligence in the following key areas at the same time: new revenue generation, process automation, risk management, customer service and customer acquisition. All AI leaders hope to adopt AI on a large scale within two years. This also confirms the following hypothesis: the application of artificial intelligence in financial services will produce significant economies of scale.
1.Current situation and development of artificial intelligence application
In the whole sample, 85% of the respondents applied artificial intelligence in some way, among which fintech companies led traditional financial institutions by a narrow margin (90% to 80%). In order to better understand the different uses of AI in financial services, this study further distinguishes AI users according to different application fields:
Generate new revenue potential
Process reengineering and automation
At present, the most common application field of artificial intelligence is risk management, followed by the formation of new revenue potential through new products and processes. However, according to the application plan and current application data statistics, artificial intelligence will be the most widely used in revenue generation in two years.
In all business areas surveyed, fintech companies are the current leaders in artificial intelligence applications. Fintech companies are ahead of traditional financial institutions in the application of artificial intelligence to form new income potential. On the contrary, traditional financial institutions occupy a higher share in the artificial intelligence being developed. Fintech companies and traditional financial institutions use AI to a similar extent in three application areas: generating new revenue potential through new products or processes (80%), customer service projects (74%) and customer acquisition (69%).
Fintech companies are different from traditional financial institutions in using artificial intelligence to realize process reengineering and automation (77% and 68% respectively) and risk management (80% and 73% respectively). However, as more mature financial services companies are developing applications or plan to apply AI in the short term, this gap may narrow.
Application statistics from different financial service sectors show that although the average application rate in the sample is the same, outliers are more common in some cases. Most notably, investment managers seem to focus on using AI to generate new revenue potential (61%), while for payment providers, this is the least active application area (44%). Similarly, the application of artificial intelligence in process reengineering and automation and customer acquisition also varies from industry to industry.
If we pay attention to a few companies that take the lead in applying AI in their core business, we will find a clear trend: all AI leaders in the survey will widely apply AI in all five fields within two years. AI leaders have obviously changed from mainly using AI to reduce costs to using AI’s ability to create new revenue, which further proves the above overall trend. 38% of AI leaders are currently applying AI in the field of income generation, which is the most active field of AI application at present.
On the other hand, the laggards of artificial intelligence seem to be far from the full application of artificial intelligence. They are particularly backward in the application of artificial intelligence to support customer service and customer acquisition. Considering the overall application gap between AI leaders and laggards, this may mean that it will not be easy to gradually shift from a simple automation use scenario to a life cycle based on AI value proposition. For example, the field of risk management can provide more accessible (or generally relevant) use scenarios for AI than process reengineering or automating complex processes.
2.Trend of large-scale application within the organization
The company is developing towards the large-scale application of artificial intelligence. A considerable number of respondents said that they are trying to apply artificial intelligence in different fields within their organization at the same time. 91% of the respondents said they hoped to see artificial intelligence applied in three or more areas of their business within two years. At present, only 42% of respondents said they had achieved this goal. According to the expectation of respondents, the “real” wide range of users with AI applications in all five fields will quadruple to 64% in two years.
This trend may be related to the scale to which AI benefits. Existing infrastructure (such as data pipeline, internal programming framework, computing resources) can be easily shared among different application scenarios within the organization. In addition, larger data sets tend to produce richer insight, and data types can also be used across different application scenarios. This is reflected in social media scenarios. For example, user insight can be used for credit analysis, while insight into online posting behavior can be used to predict stock returns.
Large scale application can also promote the company’s determination to build technology infrastructure and overcome early application obstacles. Figure 6 shows that companies that currently believe that AI is “important” or “very important” to their business are obviously applying AI more widely, and nearly three-quarters of companies are expected to apply AI in all five fields.
In general, these results show that artificial intelligence represents a series of technologies that provide basic value for financial service companies. They are applicable to many different models, and there is no necessary requirement and return for specialization. Therefore, the following chapters will elaborate on the advantages of large-scale application of artificial intelligence and potential early user advantages.
3.Specific application fields of artificial intelligence
Use artificial intelligence to create new revenue potential through new products and processes
As mentioned earlier, financial institutions can build a variety of new value propositions by using the realizable insight gained by AI from data, or by developing AI to serve other organizations.
The use scenario of using artificial intelligence to create new revenue potential mainly focuses on artificial intelligence data analysis and the use of alternative data to generate new insight. In fact, this seems to be the most widely used way of artificial intelligence in all major financial services sectors in the survey sample.
Artificial intelligence data analysis has many functions, which can find the laws in the data and connect them with business decisions. For example, MasterCard uses near real-time shopping data and artificial intelligence data analysis to generate weekly automated reports on macroeconomic trends for a variety of industries and geographic regions (mcwaters et al., 2018).
The following lists the multi sub categories and corresponding use proportion of artificial intelligence data analysis. Among all organizations that use artificial intelligence data analysis, sales analysis is the most widely used subcategory, followed by credit analysis.
4.Artificial intelligence risk management
In general, risk management is a representative field in which most entities have applied artificial intelligence. This is not only because risk management is universal as a necessary business function, but also because of the commercialization of relevant AI solutions (sweetey, 2019). From regulatory compliance to risk management or fraud detection, artificial intelligence can reduce economic costs, reduce manual intervention in micro activities, and make the risk management process faster and more efficient (Arslanian and Fischer, 2019). With the expansion of the application scale of artificial intelligence in the organization, the risk caused by artificial intelligence and the importance of artificial intelligence driven risk management may increase.
For respondents using artificial intelligence in the field of risk management, fraud / anomaly detection and monitoring is the most common application scenario, with an application rate of 75%. The effectiveness of AI in fraud detection and monitoring can be attributed to the number and frequency of transactions and the multi-dimensional / granularity of fraud patterns, which may span multiple entities, jurisdictions and industry sectors (MasterCard, 2018). The Falcon platform of FICO, a data analysis company, uses artificial intelligence driven predictive analysis to provide fraud and anti fraud detection for institutions, which is the best real case (mcwaters et al., 2018).