AI and cloud fuel bank’s new era of fast, reliable credit risk reporting

Regulatory pressure in the South African banking sector necessitates precision, accountability and speed. For a bank like Absa, with hundreds of models supporting its retail portfolio, the stakes are high. Recognizing the critical importance of early model degradation detection, Absa understood that delays could ripple far beyond the balance sheet, impacting capital reserve targets, loss forecasts, and reputation.

To stay ahead of evolving challenges, Absa reimagined its approach to model monitoring. By rebuilding the process from the ground up, the bank transitioned from a fragmented, manual effort to a streamlined, data-driven decision-making system.

Modern banking tackles legacy roadblocks

Founded more than 130 years ago, Absa now serves nearly 13 million customers across 16 countries, including 12 African nations, the UK, the US, Czech Republic and China. It offers a full spectrum of retail, business, wealth management and investment banking products.

Model monitoring is a vital part of Absa’s governance framework, underpinning the bank’s ability to manage risk and meet regulatory expectations. Over time, however, the legacy approach – built on manual scripts, siloed systems and lengthy reporting cycles – became insufficient for Absa’s complex inventory of credit risk models. The traditional process was slow, resource-intensive and costly. Highly skilled analysts spent weeks running large code batches on an on-premises platform, processing millions of rows of data from multiple sources. Generating a single model report could take two to four weeks, while developing a new monitoring framework sometimes stretched to six months or even a year before delivering results.

As data volumes grew, it became clear that this manual approach was not just inefficient; it had become a barrier to timely, strategic decision-making.

Standardized, scalable and smart

Recognizing the need for a complete overhaul of its model monitoring process, Absa established a Center of Excellence (CoE) to spearhead a new monitoring framework with clear goals:

• Standardization of reports, metrics and visuals across all models.
• Automation of manual execution to reduce human error and repetitive work.
• Scalability to rapidly onboard new models as they are needed.
• Clarity through intuitive dashboards for business and regulatory stakeholders.

The CoE partnered with SAS to move from its legacy infrastructure to SAS Viya on AWS, a modern, cloud-native analytics platform designed for performance at scale. SAS Viya has been designed to run natively on AWS, bringing the longest-standing leader in analytics to the world’s most comprehensive cloud platform.

With the capabilities of SAS Viya, the team reimagined Absa’s model monitoring process from the ground up.

• SAS 9 Content Assessment facilitated smooth migration from on-prem SAS Grid to SAS Viya.
• SAS Viya provided a cloud-ready, in-memory processing platform for rapid data processing and visualization.
• SAS Cloud Analytic Services enabled lightning-fast, in-memory processing.
• SAS Visual Analytics delivered dynamic dashboards for real-time insight.
• SAS Enterprise Session Monitor helped fine-tune cloud resource consumption.

The result was an end-to-end, cloud-based automated solution that generates reports within hours, as opposed to weeks when done manually.

Outcomes that speak volumes

For Absa, this wasn’t just a systems upgrade. It was a strategic shift in how quantitative teams operate.

Previously, analysts were primarily focused on executing models rather than managing them. With the new approach, their time is now spent analyzing results, advising the business and driving new ideas forward.

And the benefits extend far beyond efficiency:

• Accelerated reporting: Model monitoring reports are now generated automatically, reducing turnaround from weeks to hours.
• Streamlined onboarding: New frameworks and approvals are implemented more efficiently, shortening the time required to bring new models online.
• Strategic reallocation of talent: Analysts once bogged down in manual work are now focused on higher-value projects.
• Cloud elasticity via AWS: Resources scale dynamically to meet demand, with pay-as-you-go pricing that eliminates idle capacity costs.
• Built-in AI guidance: The automated “Insights” feature of SAS Viya surfaces recommendations for further optimization.

With SAS Viya on AWS, Absa transformed a once manual, routine function into an automated strategic capability that strengthens the bank’s competitive position. AI automation in the cloud now enables faster, more reliable credit decisions, improves the predictability of losses and helps ensure capital reserves meet regulatory expectations.

A platform for the future

Absa’s transformation proves that innovation isn’t only about bold new ideas – it’s about strengthening day-to-day operations through smarter strategies.

By harnessing AI and machine learning on a cloud platform built for scale, the bank is accelerating decisions and unlocking new possibilities across its business. Absa is shaping the future of banking, one smart solution at a time.