What are the most challenging aspects of working with artificial intelligence and machine learning? How do data scientists turn these technologies into opportunities for the banking industry and how did they change the way banks communicate with their customers?
The most challenging aspect is usually getting access to data in the right format. Usually, over 80% of our time is spent before we start modelling activities. This includes identifying data sources, data engineering, wrangling, and cleansing.
Data is hosted in applications, data warehouses and data lakes in order to reduce hosting costs and improve processing. Data engineering tasks are required in order to bring this information in one repository or database on a consistent format, that allow analysts and data scientists to run analytics on it.
The way our data science produces value for our customers is through our Business Insights team, responsible with setting up priorities, objectives, and guidelines, but also with making decisions based on the outputs produced by our machine learning model.
How is artificial intelligence implemented at Santander UK? Could you share with us some use cases?
Broadly speaking, there are three types of use cases that we implement at Santander UK. One of them is classification, meaning that we organise new data into existing categories based on already classified historical data (labelled). Approximately 80% of the use cases fit on this type of solution. This solution is used to classify customers for marketing purposes for example. This methodology was probably the earliest application of data science in the banking industry, and it was used to improve credit risk models.
Another one is related to forecast, where a numeric value is predicted on a specific variable based on its historical behaviour. This technique allows predicting customer behaviour and creating appropriate offers and retention programs that match customers’ needs. The methodology includes seasonality and trend analysis to forecast a future value for that variable. Examples of variables that can be forecasted using this methodology are customer balance or customer spending patterns.
Another use case is related to optimisation, which is used to increase efficiency on existing processes such as cash management. The methodology calculates all possible interactions and identifies the most efficient one.
What are the areas of Santander`s operations where data analytics has been implemented successfully?
In operations, our artificial intelligence team is making great strides in improving efficiency (eg conversational computing), reducing costs (eg automate manual processes using artificial intelligence and robotic solutions), and enhancing risk management (eg financial crime, internal audit etc).
What is on Santander`s agenda with regards to digital transformation for 2019? What are the changing trends that you foresee in the field of Data Science?
We will focus on improving and reinventing the customer experience, by anticipating the customers` needs and moving forward from a product-centric approach to one that is customer experience-based.
Artificial intelligence technology can be used in a number of ways, with anti-money laundering (AML) as one of the main areas of focus. One trend we have seen at our competitors is the use of AI to enhance risk and compliance activities such as financial crime. An example would be the implementation of anomaly detection models on anti-money laundering processes, which allow the detection of suspicious activities that traditional methodologies may miss.
You will soon be speaking at MoneyLive Summit – what topics are on the agenda and what is that you hope to achieve by speaking at the event?
Artificial intelligence will be a hot topic at MoneyLive Summit. We will discuss developments in 2019, lessons learnt so far and success stories, including use cases methodologies and we will probably make some demos to bring artificial intelligence to life.
About Angel Serrano
Angel joined Santander UK in 2017 as Head of the Data Science, focusing on implementing Machine Learning solutions to increase the value we offer to our customers, reduce costs and increase efficiencies. Prior to that, Angel worked at PricewaterhouseCoopers since 2004, both in Spain and UK, where he led one of the advanced analytics teams in the financial services sector, implementing machine learning solutions for the UK, European clients and regulators. Angel holds a BS in Physics (University of Madrid - UAM), MBA (UAH).
About Santander UK
Santander UK is a financial services provider in the UK that offers a wide range of personal and commercial financial products and services. It has brought real competition to the UK, through its innovative products for retail customers and relationship banking model for UK SMEs. On 30 June 2018, the bank has 24,200 employees. It serves around 15 million active customers, via a nationwide branch network, telephone, mobile and online banking; and 64 regional Corporate Business Centres. Santander UK is subject to the full supervision of the Financial Conduct Authority (FCA) and the Prudential Regulation Authority (PRA) in the UK. Santander UK plc customers are protected by the Financial Services Compensation Scheme (FSCS) in the UK.
Every day we send out a free e-mail with the most important headlines of the last 24 hours.
Subscribe now