South African regulators map early AI deployment across the financial sector

A joint review by South Africa’s financial authorities provides the country’s first consolidated view of how institutions are adopting AI, revealing usage patterns, investment intentions and emerging regulatory challenges.

South Africa’s financial authorities have published the country’s first coordinated overview of artificial intelligence adoption across the financial system, outlining how institutions are currently applying the technology, where investment is headed, and what gaps exist in governance and oversight.

The Financial Sector Conduct Authority (FSCA) and the Prudential Authority released the report after conducting a voluntary survey between October and December 2024. According to the publication, which drew roughly 2,100 responses from banks, insurers, asset managers, pension funds, fintechs, payment operators and lenders, 220 respondents, or just over 10 percent, indicated they already use AI in some capacity. The analysis concentrates on banking, insurance and investment firms, reflecting the segments that manage most of the sector’s assets.

AI adoption and investment levels

Banks reported the highest use of AI tools, with slightly more than half of banking respondents indicating deployed systems. Payment providers followed closely at 50 percent. Retirement funds reported more limited use, and insurers and lenders showed some of the lowest adoption rates.

Budget allocation varies significantly across segments. Nearly half of the banks already using AI projected investments above R30 million in 2024. By contrast, most insurers and investment firms planned to allocate less than R1 million, suggesting that many actors remain in pilot or exploratory phases.

Current and planned applications

Institutions indicated that today’s AI applications are concentrated in operational processes, document handling, decision support and fraud detection. Banks, lenders and payment firms also apply AI in credit scoring and underwriting.

Future ambitions include expanding real-time fraud analytics, strengthening cybersecurity detection and improving monitoring for money-laundering risks. Insurers plan to use AI more extensively for underwriting and claims administration, while retirement funds and investment firms aim to evaluate applications in portfolio modelling and risk analysis.

Generative AI still emerging

Generative AI usage remains in early stages across the sector. Current applications include document drafting, summarization and presentation preparation. Some institutions also use GenAI in marketing or client-facing content. Longer-term plans include integrating GenAI into service channels, automated reporting, customer support and workflow automation.

Risks, constraints and governance challenge

Survey participants identified data privacy and data-protection obligations under South Africa’s Protection of Personal Information Act (POPIA) as key concerns. Other risks include cybersecurity exposure, model bias, data quality limitations and the possibility of consumer harm when automated decisions are not fully transparent.

Many institutions also reported difficulty scaling AI due to talent shortages, limited technical expertise, and constraints imposed by legacy systems. Governance structures vary widely: several organisations rely on existing risk-management frameworks rather than dedicated AI oversight models. Respondents noted a need for clearer rules on accountability, human review mechanisms, model validation practices and transparency expectations.

Current regulatory landscape

South Africa currently supervises AI-driven tools through existing laws rather than AI-specific regulations. POPIA restricts fully automated decisions that have legal or significant effects unless specific conditions are met, and requires mechanisms for individuals to request human intervention. The Financial Advisory and Intermediary Services (FAIS) Act sets standards for automated advice, including oversight, testing and governance requirements.

The forthcoming Conduct of Financial Institutions (COFI) Bill is expected to introduce additional clarity on consumer protections applied to digital technologies.

What this means for the region

The survey results provide regulators, financial institutions and technology developers with a consolidated baseline for understanding how AI is entering the sector. For the region, the findings help identify where support for skills development, infrastructure and regulatory guidance may be needed. They also offer early insight into how financial services in South Africa may adapt to automation, contributing to safer, more efficient frameworks that can support the country’s broader economic development.