Our Insights

Data science and the future of banking

Using data science to drive the evolution of operations and services

Having observed how big technology companies have turned the systematic collection and normalization of user information into revenue streams, banks are embracing data science, a concept that is no longer a fringe academic discipline. Christopher Phillips, SVP, Technology Corporate Systems at RBC, spoke at SIBOS 2019 in London—via a hologram link streamed from Toronto—about the role of data science in modern day banking.

Key insights

  • Data science will help banks transition away from rules-based processes towards AI technology
  • Effective AI strategies rely on accurate and unbiased data
  • Recruiting experienced data scientists is key, and firms need to ensure these individuals are supported by effective risk managers and business experts

Data is reshaping banking practices

A number of banks are already utilizing data analytics to power nascent technologies such as robotic process automation and chatbots in order to drive business-wide efficiencies and improve customer experiences. Organizations are now looking to further capitalize on data.

Phillips told SIBOS that banks are incorporating data science algorithms in artificial intelligence (AI) software such as supervised machine learning and natural language processing to augment fraud detection and market risk monitoring. By expediting these manual processes, Phillips said compliance resources can more readily focus on other important tasks and responsibilities.

Good data equals good outcomes

If disruptive technologies are to be truly transformative, the data they rely on must be incredibly accurate and of the highest quality. Unsound data can lead to errors or misinterpretation during the AI 'interrogation' process, which could have adverse consequences. For example, data inaccuracies could result in AI-driven analytics displaying signs of bias (e.g., ideological, gender, etc.), according to an IBM research paper.1

If disruptive technologies are to be truly transformative, the data they rely on must be incredibly accurate

“Many AI systems will continue to be trained using bad data, making this an ongoing problem. Bias in AI systems mainly occurs in the data or in the algorithmic model. It is critical to develop and train these systems with data that is unbiased and to develop algorithms that can be easily trained," reads the paper.2

A further challenge in the financial services industry is that the sector can be fragmented within a complex ecosystem, meaning that data may pass through a number of systems which may impact accuracy, or the data may be unstructured.3 It is important that market participants carefully review and assess the integrity and lineage of data before it is used in any AI environment.

From a European perspective, with the recent introduction of the European Union's General Data Protection Regulation (GDPR), organizations need to be assured that their data has been acquired lawfully before it is dissected by AI software. Deficiencies in any aspect of GDPR compliance could have repercussions for companies.

A balanced approach towards hiring talent

In terms of recruitment, banks are increasingly competing against the top tier technology companies

Recruitment of individuals well versed in data science will be imperative if banks are to develop effective data-driven product suites. Phillips said banks are targeting highly talented undergraduates specializing in data science from leading universities. In terms of recruitment, banks are increasingly competing against the top tier technology companies; 80 percent of banks told McKinsey they found it difficult to recruit the right analytics talent.4

According to Helena Gottschling, Chief Human Resources Officer at RBC, "All companies need to evolve as they prepare their businesses for the future of work. This requires alignment and collaboration - between employers and employees, cross industries, and between businesses and educational institutions."5

Data science is increasingly seen as critical to bank strategies and a growing number of providers in Europe are making it compulsory for employees to be fluent in analytics if they are to be considered for promotion to senior management or leadership roles.6

While data scientists are widely sought after at many banks, it is critical that firms have a diverse pool of talent. “It is crucial that banks employ people who understand the data, but also individuals who understand the use cases for that data. It is vital there are personnel who understand the risks of data, and ensure that businesses are governed in the right way. The risk role is paramount as it stops firms doing things they should not be doing," said Phillips.

Next steps for data

Phillips also notes that these technological innovations will have a major influence on new areas over the coming months such as internal audit practices. Leading accountancy firms have acknowledged as much, with EY stating that machine learning tools could help humans analyze more information over shorter time frames, “enabling them (auditors) to use their human judgement to analyze a broader and deeper set of data and documents. It also enables them to ask better questions and to interact more with chief financial officers, audit committees, and company boards, adding value to the audit process."7

Data science is now seen as an enabler for banks' future success, and it is likely to play an increasingly ubiquitous role in strategic and business development.

You may also like

Sources

  1. IBM Many AI systems are trained using biased data
  2. Ibid.
  3. Euroclear (January 18, 2018) From data disorder to information insights
  4. McKinsey (September 2018) Smarter analytics for banks
  5. RBC (October 1, 2019) Innovation, Technology and the Future of Work: RBC Roundtable
  6. Ibid. McKinsey
  7. EY (July 20, 2018) How artificial intelligence will transform the audit