Artificial intelligence (AI) is becoming part of various activity sectors, from healthcare to law enforcement, from retail to banking. The high popularity of these systems is due to efficiency, productivity and a significant cost reduction. AI for the banking industry is not new, trading algorithms have been in place for more than 20 years. Now, the applications are extending towards improving the end consumer’s relationship with the bank.
Learning from other client-oriented sectors, banks are now investing in automated systems that help them communicate better with clients. The AI can successfully replace low-level call-center agents and can help customers with account operations, give them financial advice and even increase safety by detecting fraud tentatives.
The robots are here
Clients already use Siri and other robots as personal assistants. That is why some banks stopped spending their dollars on training more employees. This only keeps customers frustrated at the end of a phone line or behind an e-mail address. It is best to give them direct access to their humanoid help. Most of these robots are chatbots or digital assistants, but some go a step further and can even offer sound financial advice tailored to your spending habits. Back-office and fraud-prevention applications are also becoming popular.
Customer service chatbots
- MasterCard & America Express– These chatbots are embedded in Facebook Messenger for easy access. Each can give clients details about their accounts, can make some offers and provide The AmEx can even facilitate Facebook purchases.
- Barclay’s Africa– The chatbot installed for Facebook and Twitter allows clients to check balance, make payments or pay their utility bills.
- RBS– The bot named Luvo can understand natural language and acts as a simple Plans are to extend its predictive capabilities.
- Bank of Tokyo– The humanoid named Nao can offer assistance to customers in 19 languages and provides simple interaction via its inbuilt microphone and camera.
- Santander UK– SmartBank this is a voice-enabled app that doubles as a security feature. Biometrics allows customers to identify, access their accounts, pay or even report a stolen card.
Personal finance counseling
- Bank of America– The chatbot named Erica can do much more than tell you your balance or help you pay bills. It has inbuilt features of advising you on the categories where you should cut down on spending and FICO score updates.
- Capital One- The Eno chatbot offers account inquiries, payment history and a breakdown of spending.
Back-end operations
- JPMorgan Chase– The robot named COIN can handle document analysis better and much cheaper than human lawyers. It can also act as an IT administrator granting access rights based on job levels.
- Barclays Bank- Rather than freeing some time for the front-end employees, they were more concerned about checking the work of the back-end. Their robot can process loan applications, monitor risks and even trigger red flags when a potential fraud is taking place.
How will AI work in the banking sector in the future?
The previous examples show that chatbots are some of the most straightforward applications that banks can adopt. Cutting costs, automating processes and interacting with customers on a personal level are all desirable goals for the future development of AI for banking. Yet, there is more.
The blockchain is another technology that initially was the part of the cryptocurrency world, but now is extending in traditional banking. Basically, it is a public distributed log that anyone can access and check for transactions. This is a way of preventing frauds that include money laundering and identity thefts.
Finally, banks could get a bit of inspiration from Fitbit and provide you with wearables. In the future, they should leverage the increasing rise of the Internet of Things (IoT) in smart cities. This will help clients connect with banks safer, faster and in a more secure way, based on location and biometrics. The information collected by devices such as a payment bracelet can help banks learn about busy places, operation hours and spending habits.
Why should banks use AI?
Banks are all about compliance and accuracy and machines excel at providing these features in their work. While clerks become bored and tired with repetitive tasks, an AI-powered system will be equally fresh even after running continuously for 24/7. And it won’t ask for a raise either. Furthermore, it is capable of logging on all the data related to the operations and even learn from it. It will help it perform better in the months to come.
A recent study shows that one AI machine can replace eight full-time employees and three-quarters of current operations can be re-engineered, through a process called robotic process automation(RPA).
Using robots and algorithms translates into scalability. Once a system is in place and running smoothly, just replicate it as many times as it is necessary. Just imagine how long it takes for humans to learn about new products. Getting a team up to speed can take months, while in the case of machines it only takes minutes to install the necessary software.
As inferred from the examples, AI can learn about patterns and detect when users follow a good pattern. This translates into actionable advice from the application to the end user regarding managing money, being on time with payments or limiting spending habits. For the client, that means having a financial counselor on their smartphone, while for the bank, it means fewer risks.
When automating it is not only about creating digital images of current processes. It is also about evaluating the whole value chain. The art is to digitize only added value steps.
What are the obstacles and drawbacks of AI for banking?
People are very reluctant to the use of their personal data, especially financial data. Even if they know about the benefits they would collect from allowing such an “intrusion,” they are still scared about what banks use their data for. Yet, AI systems need large amounts of data for proper training. They learn much like small children or dogs: by seeing the right thing again and again. Gradually, they detect good from bad patterns by comparing what they know to the current requirement.
Also, since this was not an option so far, the banks’ data are too scattered. Some are in call centers, data repositories, and various storage methods. Not everything is in digital format, a lot is still in hard copy. It will take years before data scientist will clean and organize all that information. Only then it is ready to become a proper training set for algorithms.
No more clerks?
Current bank employees are safe, but over the next years, there will be a decline in the number of jobs offered and in the outsourcing of support processes such as customer call centers. If the pilot programs prove successful, there will be a reorientation towards using more AI-powered products instead of human employees.
Existing studies support this tendency and show that the younger generations want immediate access to information. It makes no difference if this comes from a robot or a clerk. Older generations will also adapt. Once they discover the convenience of using such systems, there is little reason to go back.
At this time, some customer requirements need to be forwarded to a human expert. As the systems learn more, they become less reliant on the supervisor and can take on more tasks, such as advising people about spending and offering contextual bonuses. The digital future of work scares many and makes pessimists fear that they will soon be replaced by a robot. Automation should be an augmentation of human capacities and a way of providing people with meaningful work, not trapping them in mindless tasks.