Digital solution providers state that one robot can work 24/7 and replace up to eight employees, without asking for days off or a raise. Also, 75% of the current banking operations can undergo robotic process automation (RPA). Techno-pessimists are alarmed, while optimists just envision ways of smoothing out the effects of what is called the fourth industrial revolution. The digital future of work can’t be reversed and will expand to every activity sector.
What started about four decades ago in gas stations with self-service pumps will become the norm in more conservative areas, including banking, law enforcement, and even government. Banking operations have been frozen in processes that have not been changed in years, but that is about to change drastically.
Benefits of using automation, robots and AI
Accuracy, predictability and removing any trace of human error are primary goals of introducing robots into the banking industry. Automated systems can ensure compliance with internal regulation every time and collect data that will be further used to calibrate the system even more.
Furthermore, such systems generate significant cost cuts after the initial set-up of the system, which can be quite expensive. Yet, the 24/7 operating schedule, low maintenance cost and, in the case of AI, the possibility of self-improvement can easily motivate the investment.
As an expert in AI solutions from Indata Labs explains, by using deep learning and anomaly detection, an AI algorithm can understand spending patterns. Since most people are creatures of habit, whenever there is a transaction that is not like the rest, either by amount, geolocation or even the language used by the browser accessing the bank, the machine triggers an alert, requesting additional verification steps from the owner. This gives clients peace of mind and saves the bank from important financial and image losses.
One of the main benefits of letting technology deal with bank processes is scalability. Once an algorithm has been trained for a set of operations, it can be replicated in countless locations and perform to the same high standards. This is a core feature when introducing new products or processes that need to be adopted by all branches in a short time.
To make full use of the benefits of automation, a bank should take a critical look at the entire value chain and not only automate processes but re-engineer first to create a simple workflow that will be afterward translated into machine operations.
The right tool for the job
The banking and financial sectors are slowly moving from the first digital age to the second. AI, cloud computing, mobile-first and digital dashboards are already the norm, and new technologies are being adopted.
Apart from RPA which is used to increase efficiency and cut costs through process automation, AI and machine learning are used for improving the relationship with the clients, increasing customization and even fraud detection.
Another tool that can prove useful in fighting crime and increasing transaction security is the blockchain approach, a framework currently popular for cryptocurrencies, but which can help traditional financial institution and state authorities to combat money laundering.
The increasing degree of smart cities and the boost of IoT is expected to help clients conduct safer transactions based on geolocation, voice and face recognition.
Challenges in introducing automation and AI in the banks
AI systems are only as good as the data used to train them and the data fed into them for calibration purposes. Therefore, getting the best to use as learning material is one of the main challenges. Currently, banks have vast amounts of data regarding their clients, operations, payment terms, credit risks and more. Unfortunately, each of these pieces of information is stored in a different silo that is not interconnected with others and almost always tributary to legacy systems. A proper AI implementation requires the centralization of data and a cleaning stage.
The second challenge is also related to data quality and focuses on unstructured data. Currently, the data which most banks use for their operations is neatly arranged in tables, but there is a wealth of information that could boost client services in e-mails, phone communication or floating around in social media. Retrieving insights from these types of documents is impossible without AI which can understand patterns and create responses. The goal is to become paperless and collect the information that previously was stored in paper archives in a digital format that is also searchable and actionable.
Some of the applications of robotics and AI that got the widest media coverage are listed below. Most of these are chatbots or digital assistants, either cloud-based or in the shape of robots and humanoids. Other applications are related to back-end operations or fraud prevention.
- Luvo (RBS) and Erica (Bank of America) - text and voice chatbots to help clients with routine operations and mobile banking;
- Nao (Bank of Tokyo) - multilingual assistant (19 languages) that also has camera and microphone capabilities;
- Smark Bank (Santander U.K.) - an app that allows clients to use natural language to manage their account and get some financial counseling regarding spending;
- COiN Contract Intelligence (JPMorgan Chase) - analyzes documents and extracts relevant clauses;
- Barclays Bank and Piraeus Bank Romania - risk monitorization, fraud detection, loan processing, drastically reducing the waiting time for applications;
- ICICI Bank - relies on robots to perform back-end operations.
Banking is catching up with the technology revolution, and in the next few years, the tendency is to invest more in automatization and AI applications instead of human employees. This effort is motivated not only by cost reductions but also by clients’ preferences. Millennials and the upcoming generations prefer to interact with technology at a time that is convenient for them.
Currently, applications are more about automating repetitive tasks and reducing business process outsourcing. In the future, when AI becomes more autonomous it could focus on core issues such as the development of new products based on customer needs, decreasing credit risks and even advising HR regarding staffing levels.