Skip to main content
Skip to footer
We're taking you to another TCS website now.


 

 

The biggest questions need the boldest answers. That's why we're using our global scale, technology expertise and collaborative spirit to move towards a better today and a brighter tomorrow together.

Ajay Atreya explores how financial institutions are leveraging artificial intelligence (AI) to establish a data-driven approach to managing controls and delivering better risk outcomes.

 

 

What are the main benefits and opportunities of AI technologies such as machine learning (ML) and natural language processing (NLP) for financial institutions in risk management?

Over the past few years, the financial services industry in Australia has seen heightened regulatory activity following The Royal Commission into Misconduct in the Banking, Superannuation and Financial Services Industry. As such, there is a huge drive to improve transparency, accountability, and customer satisfaction across the sector. To address this, financial institutions have heavily invested into recruitment, development of processes and protocols and implementation of new technologies.

Regulation will continue to broaden and deepen as public sentiment becomes less tolerant of preventable errors and inappropriate business practices. As regulatory requirements continue to grow there is increasing appetite within the sector for tech-led interventions.

The ability to define, quantify and monitor risks is significantly easier when supported by technology with the ability to process and analyse huge volumes of data. As organisations continue to invest into data transformation strategies, reliance on cloud-based solutions have grown significantly too. Its ability to process large volumes of data, at speed and with consistency, supports decision-making from forecasting to running risk models to manage assets sizes over sseveral hundred billion dollars.

Many organisations are seeking to establish a single source of truth for critical risk data elements from data definition, consolidation, validation, and consumption across different parts of the institution. The benefits of this are realised in terms of real-time availability of data for smoother regulatory reporting, insights into capital, liquidity and performance management. ML solutions can be applied to monitor data quality and enable data discovery, for example, data matching using NLP, anomaly detection using unsupervised deep learning, etc.

A recent example is that of an Australian bank, which is now offering crypto assets to some of its customers. This is a unique proposition for the Australian market and one can well appreciate the effort in risk infrastructure, processes and data analytics required to make this possible.

How can machine learning bring in efficacy for risk decision-making?

Use of big data is becoming very widespread in helping enterprises to understand consumer behaviour. The ability of machine learning to analyse, generate patterns and identify anomalies creates new opportunities for institutions to better assess and leverage their data.

Technologies can also reduce the risk of bias that human processors are susceptible to; meaning there is great opportunity to back test financial models without the same level of anticipated bias. Machine learning technologies have created new opportunities to model and stress-test scenarios with a clear understanding of outcomes based on data and not subjectivity. Applying the right set of algorithms on internal and external data helps enhance the quality of expected outcomes like credit underwriting, prediction of operational loss among others.

Further, elimination of data outliers leads to objective assessments in portfolio reviews, bringing quantitative dimensions to qualitative aspects.

Some practical applications of these include:

  • Credit Risk Model building and validation for lending units
  • ‘liveness’ detection of customers during remote KYC
  • Network analysis of sensitive customers from financial crime perspective

 

 

How can financial institutions leverage these technologies to improve their risk outcomes and comply with increasing regulation?

Let us understand this with the help of an example. The Australian Securities and Investments Commission recently released new breach reporting and related obligations guidelines, CP 340. These guidelines are subjective, rather than prescriptive, and call for financial institutions to regularly report potential breaches of core obligations and remediate, as appropriate. Putting greater onus on the banks, the guidelines call for banks to report any breaches or investigations ongoing for more than 30 days. This requires the banks to better understand, and monitor known and potential risks associated with breaches and how they are reported.

Although this may seem straightforward, it calls for changes to be made in the ways in which financial institutions’ policies are structured, the process of identifying potential breaches, where obligations could be compromised, how they are reported and how such situations would be remediated.

As such, the definition of a breach, across multiple business lines, needs to be established and translated into the required controls to prevent a breach. The end-to-end process can then be managed and supported via technology that can enable the review of individual situations and occurrences and make customised recommendations.

The use of AI allows compliance of this to be managed in a more effective manner:

  • Codification of breach scenarios and translate into executable controls; to which different categories of transactions can be mapped
  • Control environment uplift: Obligation mapping, control lineage and cognitive control validation
  • Self-learning mechanism to enhance control design based on breaches and incidents

 

(This blog is Part 1 of a 2-part series. In the next blog, Ajay will write about the how to implement these changes, the main challenges in implementing these changes, and how to overcome them. Stay tuned!)