Banks to expand their use of analytics, according to Alex Kwiatkowski, principal industry consultant at SAS’ global banking practice.
“[Banks] have not yet realised the full potential of analytics,” says Kwiatkowski. “They may have analytics in siloed areas, perhaps on individual functional areas… but the idea of having a holistic analytics platform, one single version of the truth that covers absolutely everything I think is something that is desirable but perhaps has not been fully realised everywhere yet.”
On January 13, the European Banking Authority (EBA) published a report on the use of advanced analytics and big data in the banking sector, noting that applications remain at “an early stage”.
“This may be due to institutions’ approach to new technological solutions, affected by issues relating to their legacy systems, and to adequacy of skills, expertise and knowledge, as well as by data and security concerns,” the report continued.
The aim of the report is to “share knowledge among stakeholders and, in particular, to ensure regulators and supervisors are well informed on the developments”, but Kwiatkowski is unconvinced.
“They’ve really focused on this idea that it should be about the data which is being extracted from core banking systems – effectively the structured data that banks have traditionally operated with and used for their decision making purposes, all of which is fine. However, there is also a large amount of unstructured data that is out there and that can be messages, emails, tweets, other socially generated content, it can be images, both static images, and live images from cameras and the like.”
The report outlines eight “elements of trust” financial institutions must address in their implementation and deployment of big data and advanced analytics: ethics, explainability and interpretability, fairness and avoidance of bias, traceability and auditability, data protection, data quality, security, and consumer protection.
For Kwiatkowski, there is a balance to be struck in the banking industry between adopting the eight elements while also remaining quick, agile, and ahead of the competition. He points to accusations of bias made towards the end of 2019 regarding algorithms used by Goldman Sachs and Apple Card to determine customers’ credit limits. The accusations have sparked an investigation by the New York Department of Financial Services, Reuters reported in November.
The report also explores the potential use cases for advanced analytics, such as the calculation of regulatory capital requirements.
According to a risk assessment questionnaire carried out by the EBA in spring 2019 – which was answered by 62 banks and 18 market analysts – just over 10 percent of those surveyed said they were using big data analytics for the calculation of regulatory risk requirements, while around 25 percent said the subject was being discussed.
But the report states that “from a prudential framework perspective, it is premature to consider ML an appropriate tool for determining capital requirements, taking into account the current limitations (eg ‘black-box’ issues)”.
Although the EBA’s outlook remains cautious, banks have not been just experimenting with AI and ML but are looking to use it as a competitive advantage in daily activities, according to Kwiatkowski.
“The adoption of ML and AI, whether it is in regulatory related activity, whether it is in text analytics for customer satisfaction, customer compliant resolution, all of these things are being performed in a live environment, so they are not just a lab activity, these are things that are being used with increasing regularity because why wouldn’t you. The technology is proving itself to be scalable, robust, capable of getting the right outcomes,” he says.