November 19, 2019
Table of contents
In the early days of banking, there was a strong personal connection between bankers and their customers. Now, with the incorporation of digital banking, this connection is lost, and financial institutions are searching for a way to bring it back. Coupling financial software development with AI and its analytical capabilities has a potential to re-establish this lost connection and bring a more personalized approach to finance.
The value of AI in fintech market was estimated at $5.53 billion in 2018 and is expected to reach $26.92 billion by 2024, according to Mordor Intelligence's AI in Fintech Market report. Such a big investment is justified by the fact that AI helps to lower costs, increase productivity, and contribute to customer-facing practices of fintech companies. This contribution is particularly prominent in customer service automation (including RPA in insurance and RPA in banking) churn prediction, sentiment analysis, and personalized banking.
According to studies, the majority of US millennials would happily leave their banks if presented with an alternative. As motivation for leaving, those people cited such problems as long queues and unpleasant experience with the bank’s personnel. In order to gain competitive advantage, it’s time for financial institutions to adopt AI-based customer support solutions that allow customers to solve trivial issues from the comfort of their home.
Chatbots are software incorporated into our messaging or voice applications that guide us through the accomplishment of different tasks.
Chatbots use natural language and are able to interact with customers, replacing customer support officers in particular cases. Chatbots can accomplish tasks such as informing customers, handling their complaints, and managing their accounts upon request.
For example, a bot specialized in guiding the customer through submitting insurance claims will ask the customer to submit damage pictures and will run those pictures through a fraud detection mechanism. Even though human intervention would most likely be required in the auditing process, delegating those trivial tasks to the bot still saves humans’ time and minimizes errors since people can make mistakes and become frustrated.
Chatbots rely on machine learning algorithms, so they constantly learn and become smarter and more reliable by the day.
AI-based chatbots may eventually replace humans as the first support line. According to Gartner, chatbots will be handling around 85% of customer service interactions by 2020.
Wells Fargo was the first American bank to pilot an AI-based financial assistant project. They released a chatbot operating through Facebook Messenger. Customers were able to communicate with the chatbot directly on social media through their desktop or mobile devices. They could ask this financial assistant about their spending, account balance, the location of the nearest ATM, and other trivial questions. “Very happy to help,” the assistant would say at the end of the conversation, mimicking human interaction.
AI technology allows us to take an experience that would have required our customers to navigate through several pages on our website, and turn it into a simple conversation in a chat environment. That’s a huge time-saving convenience for busy customers who are already frequent users of Messenger.
Former Head of Innovation Group at Wells Fargo
The automated financial assistant is another product of machine learning in banking. It is designed to help users navigate their financial plans of saving and spending. Financial assistants can be used in insurance plans (such as pension or child benefits), dividend management, product discovery, and tax advice among others. Of late, more and more neo-banks are applying AI in wealth management, offering their customers robo-advisory platforms and other solutions to analyze the stock market with machine learning while automating financial planning and investment management. This kind of support increases user engagement with the financial product as people are more likely to disclose sensitive financial information to a robot than to a human from the fear of embarrassment.
One such example of a financial plan guide is Ella of Sun Life Assurance Company (Canada). Another example is Ceba, a financial assistant offered by Commonwealth Bank (Australia) that assists customers with more than 200 financial tasks.
As a result of processing big amounts of financial and nonfinancial data, AI allows financial providers to anticipate customer needs without customers needing to act themselves and provide personalized banking experience much easier.
Incorporating AI in everyday tasks helps in identifying debtors, thereby speeding financial service processes. Advanced text mining can also be used to analyze customers’ digital footprint on social media or elsewhere on the internet. This is particularly helpful in countries where customer credit history is not stored.
AI is also used to categorize people based on their credit scores. This information helps fintech companies decide how different financial services offered by the company correspond to customer categories.
It is important for financial institutions to follow the concept of a 1 to 1, personalized relationship. Traditional campaigns, such as bulk emailing and messaging of some irrelevant general texts (such as “do you want to buy our product of the month?”), will not evoke the reaction you are waiting for. Studies by The Financial Brand demonstrate that customer response rates for such campaigns are less than 1%.
People are looking for something more personal. This is achieved by a combination of decision making and a real-time data analytics strategy spanning different channels (even the ones that you do not own). You can use that to define your most value-adding action.
Here, AI can generate personalized email content and campaigns that are aligned with a customer’s past behavior. This means a lower spending as the message will be more targeted and relevant.
AI enables smart interactions: insights derived from the AI analysis of customer history can be used by the staff to pleasantly surprise the customer during their interaction with a tailored service.
Sentiment analysis (also known as opinion mining) is used to extract opinions and understand emotions embedded within a text.
Sentiment analysis tools search through texts posted by your customers on review sites, forums, social media, and other online platforms. They incorporate cognitive technologies to comprehend the nuances of human language and are capable of understanding if the tone is positive, negative, or neutral.
They can also understand the way people express themselves on social media, including slang and sarcasm, and recognize if the message aims to convey an opposite emotion or if there are some cultural references involved. Those tools can detect emojis that often stand for clouded sarcasm. Additionally, some sentiment analysis tools can work with audio and video.
With sentiment analysis, you can see how customers feel about your company, which helps you spot any wrong doings on your side, improve your customer experience and acquisition in general.
The Bank of Italy involved AI to understand customers’ sentiment for five big European banks. It analyzed Twitter feeds mentioning any or all of the five banks, in order to gain a better understanding on how the customers perceive each one of them. This analysis helped The Bank of Italy to understand customers’ general preferences and where the bank itself stood among competitors.
During the Christmas season, Expedia decided to run a new marketing campaign in Canada. They opted to include violin screeching in the background of their campaign. Annoyed customers took their frustration to social media and other platforms. With the help of sentiment analysis, Expedia was able to notice this ‘flame’ before it went too far and took corrective actions. They removed the ad and posted a series of follow-up videos including the one of the original actor smashing that unfortunate violin.
Customer churn is an important KPI for every industry and according to the study by Bain and Company, the cost of acquiring a new client can supersede the cost of retaining one by as much as 700%. Therefore, it is essential to take control over the issue instead of going with the flow.
There are some retroactive approaches to customer churn, such as A/B testing of the user interface. Yet these approaches are too slow and do not deliver timely results.
Fortunately, incorporating AI allows predicting and therefore attempting to control customer churn. AI analyzes customer behavior and presents a list of those who are likely to stop using your company’s services. An indicative behavior can be unsubscribing from newsletters and mailing lists, viewing account policies, etc. This gives managers a heads up so that they can act on the situation by improving the offer. This can potentially change a customer’s intention to leave the company.
Nowadays, the customer is in the driving seat. If you are not willing to offer them what they need, they will not have a problem leaving you for a competitor. The good news is that AI provides an excellent opportunity for banking and financial companies to shine and offer unique services to their customers. Those AI applications can either touch the customer directly, or occur behind the scene (such as sentiment analysis and churn prediction).
AI applications that touch the customer directly can include 24/7 customer support offered by a friendly and knowledgeable chatbot instead of a call center operator who can get frustrated at the end of the working day.
Another example of AI application helping you win the hearts of your customers is surprising them with a specifically personalized offer based on their social media posts. This can show your customers that you really care. However, ensure that there is a balance between ‘we care for you’ and ‘we constantly watch you’.
AI in fintech is a relatively new field, and as such, it still has its challenges. Some are associated with potential errors of the technology itself. Some involve malicious intents. For example, hackers can use AI for large-scale attacks targeting financial apps. Another example is malicious intent by employees working with AI if they disregard data privacy and ethics. Finally, there are challenges related to potentially replacing workforce with AI, at least to some extent.
However, these challenges can be overcome with a meticulous approach to data privacy, security, and bug-proof software engineering—an approach ensured by mature fintech development vendors like Itransition.
We help financial market players and enterprises of all sizes build custom financial software solutions for internal and customer-facing business processes.
Find out why you need to prioritize customer service automation and learn where to start your automation journey.
Why not exploit your big data with predictive analytics to gain the insights for better fiscal decisions? This article gives all the answers.
In the face of the growing uncertainty surrounding machine learning adoption in the banking industry, Itransition responds with an actionable framework.
Find out about Itransition's high-profile BI consulting for a Canadian bank, including data architecture analysis and a BI strategy.