The World Economic Forum’s report on the future of the labor market estimates that in 15 countries accounting for 2/3 of global jobs, 7.1 million jobs will be eliminated in the next five years, and 2 million new jobs will be created. This clearly negative result will be partly due to the development of automation and robotics. How does this affect the banking sector?
The scenarios we have up until now known from science fiction films are becoming reality: robots are being successfully created and used by world giants and not only. Financial consulting seemed to be a field that was immune to automation, as contact with a competent, “living” person seems invaluable in such a matter as sensitive as personal finance. However, fintechs have appeared even in this area, and, using artificial intelligence, offer access to personalized advice. Moreover, they promise that it may be cheaper and decidedly more widespread.
The first intelligent solutions in the field of personal finance primarily hit the property management segment. Properly designed algorithms help those who do not want to make difficult decisions themselves with investments. Many clients are ready to give up contact with humans in return for lower fees and access to services that have so far been reserved only for wealthiest members of society. For a fraction of the amount usually charged by investment advisors, automated advisers are ready to diagnose a client’s profile and then offer him the appropriate securities and periodically adjust the portfolio to changes occurring on the market. Real financial advice does not have to be a service available only to the elites anymore. This market has been flourishing in the United States for several years.
Scheme of action
The scheme of action of AI advisors resembles the cooperation with a “living” financial adviser in many respects. First, the client responds to a number of questions about his financial situation, the products he owns, and his plans and goals.
The second step is the estimation of the level of risk aversion. Forms in which the client indicates how he would react to specific situations, such as the loss of a portion of his capital, are typically used. Some solutions take into account the tendency of people to overestimate their “appetite” for risk and then correct their answers.
The process of “getting acquainted” with the client is mainly the presentation of a portfolio proposal as well as the presentation of the simulation. The client may choose to entrust the management of his capital to the investment advisor, who will then make the purchase of the appropriate assets on his behalf. The funds go to the brokerage house cooperating with the advisor.
Artificial Intelligence in Banking
The times when banks had huge revenues from loans and credits, and regulations were not particularly strict, have passed. Further regulations have restricted the whole industry, and new competitors, such as the PayPal payment platform and investment funds such as Wealthfront, are robbing banks of clients they did not have to fight for until recently.
Artificial intelligence could turn out to be what banks need to improve their position among the new fintech competitors. In a sector where contact with people is crucial, major changes in customer service approach have been taking place.
In the United States there are already companies providing financial intelligence services based on artificial intelligence. A self-teaching algorithm based on the neural network schema is intended to help the client with the overall planning of his personal finances. Experts estimate that the cost of such service does not exceed 1/20 of the rate proposed by professionals.
Aggregation of accounts, financial goals and forecasts
Supporting artificial intelligence is not possible without a large amount of data. Therefore, service providers implementing AI solutions start by proposing the integration of the customer profile with his bank accounts. The aggregation of accounts is based on a system of collecting data from bank accounts, brokerage accounts and credit providers. The algorithm analyzes the client’s expenses and, based on his actual behaviors (rather than his declarations), defines, among others, the household savings rate, fixed expenditures, and other indicators relevant for financial planning. Artificial intelligence constantly monitors changes in client behavior and can react to new events, such as revenue growth.
The next step in building a financial plan is setting the goals that the client intends to realize. These goals can be very traditional ones, such as the raising and education of a child, the purchase of real estate, or preparing for retirement, but can also be a change of career path or year off from work.
In each case, the algorithm estimates the costs associated with achieving the goal, both one-off and repetitive, and suggests funding sources. For example, in the case of a property purchase, AI advises the client on what kind of property he can afford in a given area, what financial burden the mortgage will put on him, and the formalities involved in the transaction, taxes, and property maintenance. Artificial intelligence also indicates which of the planned ventures are unrealistic in the given situation and what steps need to be taken to change this.
Getting millenials with chatbots
Chatbots are, according to some experts, the future of consumer banking. There is a bot that communicates with users using colloquial language, meaning it can read emotions, understand emojis (!), and can even be witty. Chatbot Eno (implemented at Capital One, an American bank holding company), because that is its name, is likely to be widely used in banking – especially in countries where mobile applications have not faired very well yet. Why? Banks install chatbots because they know that not only will they reduce the costs of customer service, but will also allow them to gather new information about their customers.
Eno is a gender neutral bot. Its basic functions are the ability to answer user questions, which are often asked using colloquial language. One does not need to formulate questions concerning account balance using the words “balance” or “checking account”, but rather simply ask: “how much money is left on my card?”. In addition, with Eno, one can perform simple banking operations such as payment for the account or transfers to trusted recipients. The bot can detect a sense of humor and possesses the ability to use emojis.
The Indian HDFC Bank, whose Eva (Electronic virtual assistant) bot has similar functions to Eno, is also experimenting with this technology. Just several days after its implementation, it responded to more than 100,000 questions from this bank’s clients.
Thanks to Eva or Eno, the bank not only knows what the most frequently asked questions are, but also when, where, and under what circumstances they are asked. It also possesses extensive knowledge on a user’s social media. The effective use of this information allows chatbot owners to profile more effective ads.
In March 2016, another Capital One chatbot, the voice-activated Alexa, entered the market. Unlike other types of chatbots that you can use with your smartphone or computer, Alexa requires the purchase of additional equipment, something which could slow down the process of its dissemination. Despite this, Mark Sullivan of Fast Company, an American technology magazine, thinks Alexa will be the standard in home-used AI.
It allows its users to take advantage of a variety of services: checking accounts, credit cards, or applying for a loan. Unlike mobile applications or web chatbots, Alexa can be updated at any moment, without the client needing to download the updates himself. Analysis shows that the most important thing for Alexa’s clients is to check their account information, not obtain information about the products or services offered by the bank. Some of the most popular questions are “What is my account balance?” and “How much is on my account?”. According to the creators, focusing on the specific types of questions and perfecting the scripts used to respond to them have largely contributed to its success.
As with any new technology, unexpected results have emerged, which is why voice bots are developing so dynamically. In Messenger, it has been possible to enter text verbally for a long time. The application is especially noteworthy because it is relatively good even with difficult languages like Polish. For example, imaginBank’s (the Spanish Caixa Bank’s) chatbot answers questions about loan offers as well as store discounts and rebates in Facebook Messenger.
Twenty per cent of Europeans use Messenger in their daily communication. Chatbots, like those installed on this application, are likely to revolutionize consumer banking.
Should “living” advisors be afraid of AI solutions?
Artificial intelligence is to be much cheaper. Than human advisors With low fees, bots can reach people outside of the circle of clients of traditional advisors. An advantage of the algorithm’s constant “care” will definitely be the constant updating of the data used to prepare forecasts. One of the biggest concerns connected with the development of automation and robotics is the risk of a rise in unemployment. However, Elon Musk, the owner of Tesla, an innovator and supporter of artificial intelligence, said that this problem could be solved – “With automation, people will be able to address more complex issues.”