The banking and financial sectors are on the brink of a fundamental structural change. Compliance regulations have kept these industries trapped in long-overdue legacy systems. Yet, senior executives understood that they need to make a shift towards integrating
more technology into their daily operations.
This decision is motivated by costs, efficiency, perception, and evolution. The current buzzword which is credited to be a catalyst for the financial sector’s revolution is AI. This is due to the pressure posed by Fintech start-ups on traditional banks and
a sharp decline in the latter’s credibility among the new generation.
Financial Predictions and Deep Insights
Fintech is mostly based on data and using the insights to get ahead of competitors. As long as the inputs are high-quality, clean, and relevant, the results can create scenarios, build portfolios and anticipate future trends accurately. There is a real concern
related to the quality of the data used by the algorithms. These work as black boxes and have no corrective mechanisms, therefore, it is a matter of garbage in, garbage out.
Through the democratization of AI tools, even small companies or individuals will be able to compete with global giants. A key area in fintech is predicting the customer’s behavior. Whether the company wants to know if the client is inclined to make another
purchase or if they will repay their loan without missing payment, such answers can be found in data science.
The downside is that most algorithms work like black boxes and they cannot provide the reasoning behind the decision. Yet, this problem can be overcome by giving the algorithms a support role, in the hands of a specialist.
Fraud Detection and Prevention
The financial sector is one of the primary targets of attempted fraud, and this is due to increased security breaches of online transactions. A solution to this is to use a
Generative Adversarial Network (GAN) which replicates data with dummy content to teach the neural network to spot the difference between real data and the hacked one.
Anti-fraud algorithms aim to identify any transactions that don’t follow the usual patterns, like accessing an account from the same device or the same location. Not all these are clear indicators of a threat, as a customer could be traveling, for example.
The role of a deep learning algorithm is to go over numerous transactions and try to identify the patterns behind those that were confirmed as frauds. Through this process, the algorithm becomes more refined, and it can fight potential future attacks too.
Other ways to ensure security is to use image identification algorithms. For example, AI could get scanned documents as input and verify authenticity. Fingerprint verification or even a selfie log-in can act as additional security factors for accounts.
AI and deep learning algorithms have been used in trading for the last 10 years, but only recently has the technology begun to play a more active role in the direct relationship with clients. Both start-ups and banks are creating and deploying chatbots to
consolidate their relationship with clients and to free their staff’s time.
Some of these tools could actually replace call center agents. This is due to millennials’ preference for a quick answer, preferably by text, instead of calling. These tools appeal to the new generation through their transparency and ease of use. Also, talking
to a chatbot means a high degree of personalization.
Some pioneers in the chatbot field include Amexbot from American Express which is a dedicated account manager for each client and ERICA from Bank of America which even has inbuilt predictive analytics to help clients manage their budgets.
AI and Automation
There is currently confusion in the world of fintech. Not all software is supported by actual AI, and most applications have to do more with automation. The difference between the two boils down to problem-solving versus optimization.
Most companies don’t use AI, they only use data-intensive algorithms that provide some degree of automation for their usual processes. This is just what the production world had been doing for years, taking the latest technological advancements and increasing
the efficiency of repetitive tasks. Examples of this type of work include scoring loan applications and answering simple client inquiries.
True AI aims to replicate cognitive processes of highly skilled staff. The work of AI could potentially replace those of investment brokers, financial valuators and loan evaluators. The goal is to make accurate future predictions, create scenarios and take
Concerns About AI in Fintech
Although using AI promises some benefits, including cost reduction, efficiency, improved security and decision-making, the technology is too young to be fully understood and tested. The financial sector is especially vulnerable to attacks or misconducts.
While legacy systems developed as a response to past problems made the operations cumbersome, the risks of fintech are not even wholly mapped yet. In fact, a survey shows that 76% of financial professionals think that regulators are not familiar with the
risks of AI.
A relevant example comes from an
incident involving the Associated Press. In 2013, their Twitter account was hacked and a message that President Obama was killed was released to the public. This simple action had devastating effects on the automated financial markets. Text analysis algorithms
picked up the news from a trusted source and changed their behaviors accordingly, without verifying the information. The damage was estimated at $136 billion in three minutes.
Is AI Outgrowing the Hype Phase?
Until now, AI seemed like another fad, but it is beginning to emerge victorious from an experimental phase, towards mainstream acceptance. AI can generate an almost unfail competitive
advantage if used correctly. This advantage could be best seized by small and medium companies in their competition with industry giants.
In fintech, it will most likely not replace humans any time soon, but it will offer them excellent tools. The applications are numerous both for financial sector workers who will have AI-powered decision tools and for end users who will enjoy robot-advisors
and fraud prevention directly on their smartphones.