Accelerating stress testing in the cloud (2024)

Intesa Sanpaolo enhances efficiency and meets stress-testing requirements six times faster with SAS^® Viya^®

European banks are navigating an increasingly complex regulatory environment, where rigorous stress testing and risk management are vital to ensuring stability and compliance. Within this context, Intesa Sanpaolo, Italy’s leading bank group and a prominent member of the European banking community, recognizes the crucial role of digital transformation.

That was the impetus for Itesa Sanpaolo’s Democratic Datalab project and the bank’s move to SASViyaon Google Cloud. The technology made it possible for Intesa Sanpaolo to more quickly and intensively complete its most recent European Central Bank stress tests.

“We finished the stress test six times faster than before by rewriting some of the underlying models and using SAS Viya, a more efficient and cloud-accessible analytical platform,” says Alessandro Ruffa, Head of Platform Financial, Enterprise and ESG Risks at Intesa Sanpaolo. Ruffa’s organization is responsible for developing data-driven digital services within an open, collaborative model such as the one supported by SAS Viya and the Democratic Datalab.

In this Q&A with Ruffa, we explore Intesa Sanpaolo’s digital transformation, the motivations behind the bank’s pioneering Democratic Datalab project, its early outcomes, and Ruffa’s insights on artificial intelligence in risk management and banking.

What was your path to SAS Viya and approach to advanced analytics issues?

Alessandro Ruffa: The Democratic Datalab project marks the beginning of our journey to digital transformation, emphasizing open innovation and fostering a democratic, collaborative work environment. The initiative aims to enhance communication, operations and productivity by bridging the gap between our IT and business departments.

We had been using SAS on-premises for several years, and all our analytics models – especially those dedicated to risk management – were written in SAS. Through the Democratic Datalab project, we developed a new technology platform that uses sandboxes and predictive models to rapidly create and test digital solutions and services that are ready for deployment, thanks to the collaborative work of our fusion teams.

Fusion teams are a blend of technical and business experts using a hybrid approach of Waterfall and Agile methodologies. The fusion teams approach fosters early involvement from technical and business stakeholders, encourages skill exchange and accelerates service delivery while ensuring the final solutions meet our operational needs.

After our move to Google Cloud and SAS Viya, we made significant advancements by rewriting and optimizing some of the underlying mathematical and SAS models. The performance of Google Cloud and the openness of SAS allowed us to use modern programming languages, including Python and R, which helped us support our data scientists more effectively.

This also streamlined our stress-testing processes, dramatically reducing execution time to one-sixth of our previous time and improving performance due to the fusion team’s collaboration. The fusion team included around 20 people from the IT and business departments who developed the stress test, plus six to seven end users who conducted the tests, showcasing the power of close collaboration between the IT and business functions.

How does your collaboration with SAS, including access to its resources and expertise, integrate into your operations? What led to the partnership, and what are its key strengths?

Ruffa: We have worked closely with SAS experts to optimize the performance of the models running on the platform. SAS Viya’s ability to support modern programming languages has been especially beneficial, enabling us to onboard new talent who are trained in Python and R – this is a great hiring and project accelerator.

And the human aspect has been vital. The SAS professionals we’ve collaborated with understand our organization and have extensive experience in our specific area, which is crucial for our critical design and transformation initiatives.

In addition, we operate in a sector that is regulated by very stringent, constantly evolving regulations. We’ve found SAS’ professionals to be exceptionally knowledgeable about those regulations. Their expertise has ensured we’re always compliant – and often ahead of regulatory changes – and supported by a technology platform that quickly adapts to new requirements.

AI algorithms are increasingly part of our daily lives, but can we trust AI? What’s your perspective?

Ruffa: For the time being, we have not implemented AI internally; however, we are exploring its potential. I believe AI can offer enormous benefits in certain areas, especially with task-specific algorithms.

For example, we are experimenting with generative AI to assess how well our internal policies align with upcoming regulations and determine the impact of these new rules on our policies and processes. These analyses, often very time consuming, can be efficiently supported by AI, provided that the final evaluation and decision making are not left entirely to the systems and are overseen with proper governance and control.

So my perspective on experimenting with and adopting AI is certainly positive. However, I believe we should ensure the technology has adequate human oversight with a dedicated team, committee or board that ensures governance – monitoring, verifying, managing and making the final decisions.

In a future where AI brings greater efficiency, how do you see performance and productivity evolving in your organization?

Ruffa: I can answer that with one word – simplification.

In a complex and highly regulated organization such as ours, processes must be strictly controlled to maintain governance, so aiming for improved performance may seem unrealistic.

However, even in such contexts, dynamism and flexibility (which I would use as the very synonyms for performance and productivity) can be achieved by simplifying processes and the way processes are managed and implemented. And here technology plays a key role. In our Democratic Datalab project, the dynamism and agility of the collaborative processes are only possible because of the underlying technology.

The results are evident in real performance and productivity numbers: Stress tests were completed six times faster, even with more intensive and accurate testing.

Conclusion

Intesa Sanpaolo’s digital transformation streamlined the bank’s response to regulatory demands and positioned it for future advancements in AI, promising more precise risk management and operational resilience. This strategic embrace of digital innovation, cloud computing and advanced analytics platforms has marked a crucial step in addressing the evolving challenges of risk in European banking, ensuring Intesa Sanpaolo remains competitive and compliant in a rapidly changing landscape.