Banking 3.0: Intelligent Services and Financial Health

Customers are demanding innovation from traditional financial services and expect banks to provide intelligent and customized services that offer added value, all via the fastest and most efficient channels to secure an outstanding user experience. Financial institutions have access to vast quantities of data on their customers: the exact time that each transaction is performed, transaction amounts, where the purchase was made and how frequently customers use specific channels…

The large number of financial movements generated by customers each and every day create an ideal scenario for harnessing data science, understood as synergies between machine learning, data mining, big data and statistics, to infer insight and offer added-value intelligent services to customers. It is also the perfect scenario for making the leap to banking 3.0. So what does this have to offer?

What customers will notice most about banking 3.0 is how they interact with their “intelligent accounts” and services. Customers can communicate with their accounts using their own normal voices, holding a conversation in natural language without even having to touch their mobile device screen or click on the bank’s website. Humans are accustomed to using natural language and it is how they communicate most comfortably. Which is why banking 3.0 allows customers to interact with intelligent and personalized services using their natural voices, in the manner of a conversation. The added-value banking 3.0 services are described below:

Intelligent account services

1. Predicting everyday expenses: predicting customers’ regular bills, such as electricity, water, mortgage, insurance, telephone/internet, and so on. Customers can use natural language to ask their account “how much will the next water bill be?” or “how much more will I spend on electricity in 2018?”  [1].
2. Predicting overdrafts: [2] predicting account movements in advance, helping to detect potential overdrafts and warn customers before they occur, thus helping them to make sound decisions. Customers can ask their account which month they will be in the red. A potential overdraft can be predicted well in advance, with the account then offering the customer a tailored financial product, all as part of a dialog using natural language. The customer can in turn use his/her voice to arrange the appropriate financial product to avoid the overdraft.
3. Evaluation and notification of deviations against anticipated spending: the ability to predict movements means customers are able evaluate their own spending behavior when significant deviations arise between predicted and real spending levels, with a particular focus on regular living expenses [3]. Customers can then be warned of any unsustainable spending patterns, service issues, or if they have contracted too much power from electricity service providers.
4. Automatic sorting of movements: Every debit or credit movement made to the account will be automatically organized into categories, allowing for innovative analysis from an entirely new perspective [4]. Automatic categorization means customers can benefit from real-time ‘financial health‘ monitoring, including quick analysis of income and expenditure. Customers can use their voice to find out how much they have spent on clothes this month or how much they have spent at their favorite restaurant over the last three months. The intelligent account will then provide a response in natural language.
5. Customers can evaluate their own spending patterns: they can independently evaluate their own movements, meaning they will be able to take action to improve their spending habits and decision making [5]. Customers will be able to ask their account if they have already paid their half-yearly home insurance premium. The account will then answer in a conversational format, “yes, the premium has already been paid”, stating the date and amount of the payment, or if it is still outstanding the customer will be told when the payment is set to go through and how much it will be.
6. Comparing spending patterns against anonymous customers with similar profiles: we often ask our friends and family how much they spend on regular living expenses to compare against what we spend at home [6]. This service provides customers with a benchmark to help them identify whether their own household spending is in line with similar customers. The comparison is based on aggregated and anonymous data from customers with similar profiles, all in real time. The intelligent account will inform customers, using natural language, whether specific spending on a regular service is in line with the average for similar customer profiles. A customer with a similar profile is one in the same income range, with the same family size, a comparably sized home and in the same city.
7. Recommendations based on account movements: by evaluating the customer’s daily movements, the intelligent account will recommend products that could improve their experience and satisfaction with the bank, providing a fully customized service in natural conversational language. For example, if a customer does not have much income and is always making debit purchases, the account would recommend a credit card, which can then be arranged quickly and easily using natural conversational language [7]. The service may also include analysis of the customer’s social media posts, provided this is public and not private information, to then offer recommendations. For example, should the user post social media photos and comments about a car they want to buy, the account might offer a customized credit facility to finance the purchase, which can then be arranged conveniently using natural language.

Porras Castaño, Javier

Expert in Innovation, Data Scientist, and Doctoral Student

Bibliography

1. Victor I. Chang, Muthu Ramachandran: Financial Modeling and Prediction as a Service. J. Grid Comput. 15(2): 177-195 (2017)
2. Stanislav Sobolevsky, Emanuele Massaro, Iva Bojic, Juan Murillo Arias, Carlo Ratti: Predicting regional economic indices using big data of individual bank card transactions. BigData: 1313-1318 (2017)
3. Sandra Mitrovic, Gaurav Singh: Predicting Branch Visits and Upselling using Temporal Banking Data. CoRR abs/1607.06123 (2016)
4. Mauro Castelli, Luca Manzoni, Ales Popovic: An Artificial Intelligence System to Predict Quality of Service in Banking Organizations. Comp. Int. and Neurosc. 2016: 9139380:1-9139380:7 (2016)
5. Mauro Castelli, Luca Manzoni, Ales Popovic: An Artificial Intelligence System to Predict Quality of Service in Banking Organizations. Comp. Int. and Neurosc. 2016: 9139380:1-9139380:7 (2016)
6. Binoy B. Nair, V. P. Mohandas: Artificial intelligence applications in financial forecasting – a survey and some empirical results. Intelligent Decision Technologies 9(2): 99-140 (2015)
7. Ivens Portugal, Paulo S. C. Alencar, Donald D. Cowan: The use of machine learning algorithms in recommender systems: A systematic review. Expert Syst. Appl. 97: 205-227 (2018)

References

1. EVO Bank. It’s not magic, it’s Big Data (click here)
2. Expansión: Spanish banking puts its future in the hands of ‘big data’ (click here)
3. BBVA: What financial ‘big data’ reveals and how it is used at BBVA (click here)
4. Álvaro Martín BBVA Research. What is artificial intelligence? (click here)
5. El País. Retina. Social challenges and opportunities in artificial intelligence (click here)
6. Universidad Carlos III de Madrid. Artificial intelligence to improve data processing (click here)
7. CSIC. Artificial intelligence and recommendation systems (click here)
8. Cinco Días. From big data to the smart company (click here)
9. TELOS. Telefónica. Living in a sea of data. Artificial intelligence (click here)
10. Universidad Carlos III: Using artificial intelligence to identify behavior (click here)