Use of artificial intelligence in the financial sector: opportunities, risks and future developments

The rapid development of artificial intelligence (AI) in recent years has had a profound impact on many sectors of the economy. Especially since the breakthrough of generative AI applications like ChatGPT, AI has captured public attention well beyond tech circles. In Switzerland’s financial sector, AI offers enormous potential to streamline business processes, improve decision-making, and boost efficiency.

Urgent Azizi  Head of IT Audit & Advisory in the Financial Services practice, BDO Switzerland

However, the adoption of AI also brings legal and regulatory challenges. The complexity and often opaque workings of AI systems create uncertainty for financial institutions. There are still concerns about data protection and security. These issues raise important questions about how AI will shape the sector’s future.

Many Swiss financial institutions are already integrating AI into daily operations—sometimes intentionally, sometimes without realizing it. FINMA’s recent survey shows most institutions initially test AI in low-risk areas. Nevertheless, some have progressed to more advanced applications backed by structured risk management.

Use cases for AI in financial institutions

Currently, a number of central use cases in financial institutions can be identified in which AI contributes significantly in optimizing and securing business processes.

AI-powered transaction monitoring, which uses state-of-the-art technology to detect risky transactions. Algorithms are used that are able to identify new, previously unknown patterns that conventional, rule-based systems often overlook.

AI is also playing an increasingly important role in cyber security. Advanced AI algorithms help spot and react to cyber threats earlier, enhancing institutions' ability to defend against evolving attacks.

In the area of customer interaction, AI-driven chatbots allow customers to manage accounts, block cards, check credit limits, and filter transactions—cutting manual work significantly.

AI is also used in areas where straightforward efficiency gains can be achieved. AI supports functions like document translation, text generation, and automated documentation, freeing up human resources.

Finally, there is also a trend towards AI-supported financial advisors. AI-based advisors analyse vast datasets in real time to offer personalized investment strategies, assess risks, and continuously optimize client portfolios.

These developments underline how AI technologies are acting as a catalyst for innovation in the financial sector while helping to improve security, efficiency and customer satisfaction.

Regulatory approach in Switzerland

The regulatory approach in Switzerland, as formulated in the AI guidelines of the Federal Office of Communications, aims to foster responsible and transparent use of AI without stifling innovation. The guidelines emphasise the need for a balanced framework that supports technological developments while taking ethical principles, data protection and security aspects into account. In particular, AI systems must be implemented transparently and securely in order to strengthen the trust of users and market participants. Switzerland plans to integrate the Council of Europe’s AI Convention into national law, ensuring harmonized and targeted regulations. Legal adjustments should be made as sector-specific as possible, so that general, cross-sector regulation remains limited to central areas relevant to fundamental rights, such as data privacy. In addition to legislation, non-legally binding measures are also being developed that can be implemented in the form of self-declaration agreements or sector-specific solutions, for example.

The approach is based on the following four key studies: International regulatory comparisons across 20 countries; Legal analysis of the AI Convention, the EU AI Act, and Swiss law; Sector-specific assessments for federal law amendments, and an economic and policy impact evaluations.

FINMA’s expectations

In its Supervisory Guidance 08/2024, FINMA stresses that financial institutions must approach AI adoption systematically and risk aware. Although Switzerland has no specific AI regulation, existing financial market principles apply. Institutions must establish a clear governance, systematically inventories AI applications and define responsibilities. In addition to data protection risks, FINMA particularly emphasises the challenges in the areas of model robustness, bias and explainability.

A key concern is the quality of the data used, as incorrect or outdated information can pose unexpected risks. FINMA is therefore calling for clear guidelines on data control and regular tests to ensure the accuracy and stability of AI models. It is also critical of the potential for data drift and inadequate model monitoring.

Institutions must ensure traceability and be able to explain AI decisions in such a way that they can be understood by customers, employees and supervisory authorities. In addition, an independent review of the models is required to identify and reduce model risks.

FINMA also made it clear it will refine its expectations in line with international standards.

Recommendations for action

Going forward, the targeted and strategic use of AI will be a decisive factor for success in Switzerland’s financial sector. Institutions must embrace AI opportunities to remain competitive while ensuring that they establish robust governance, risk management, and compliance structures.

Across the industry, regulatory bodies and academia agree: AI will bring sweeping changes that cannot be ignored. Institutions must proactively engage with AI, weigh its risks and benefits carefully, and integrate it thoughtfully into their business strategies. By doing so, institutions can ensure AI becomes a long-term asset and not a liability.

This article was created with the support of artificial intelligence. The information is based on our market perspective as well as publicly available sources and regulatory requirements.

Biography

Urgent Azizi has been working as a consultant for IT and operational risks for over a decade, with a clear focus on information security, IT risk, and strengthening resilience in the financial sector. Throughout his career, he has supported numerous financial institutions including Global Systemically Important Banks (G-SIBs) and systemically relevant banks in Switzerland in implementing and assessing regulatory requirements in the areas of IT, information security, and resilience. Urgent Azizi is currently Head of IT Audit & Advisory in the Financial Services practice at BDO Switzerland. In this role, he supports clients both strategically and operationally in addressing complex challenges related to cyber security, IT compliance, operational risks and digital resilience. 

List of sources

• Swiss Financial Market Supervisory Authority (FINMA) (2024). Regulatory Notice 08/2024 – Governance and risk management in the use of artificial intelligence. Berne: FINMA.
• Federal Office of Communications (OFCOM) (2024). Guidelines for the use of artificial intelligence in Switzerland. Berne: OFCOM.
• Council of Europe (2024). AI Convention – Legal framework for the use of artificial intelligence in Europe. Strasbourg: Council of Europe.
• European Commission (2023). Artificial Intelligence Act (AI Act) – Regulation of artificial intelligence in the EU. Brussels: European Commission.
• Financial Stability Board (FSB) (2024). The Financial Stability Implications of Artificial Intelligence. Basel: Bank for International Settlements (BIS).
• OECD (2024). Explanatory Memorandum on the Updated OECD Definition of an AI System, OECD Artificial Intelligence Papers, No. 8. Paris: OECD Publishing.