Serbia’s financial sector is presenting two very different pictures of modernisation.
On the conventional measures of resilience — capital, liquidity, profitability and credit quality — institutions are becoming stronger.
At the same time, adoption of artificial intelligence within core risk-management processes remains relatively immature.
Recent industry assessments put AI maturity in Serbian banking risk functions at only around 28%.
That gap is significant.
Serbian banks have undergone extensive digitalisation on the customer-facing side. Mobile banking, electronic payments and automated consumer processes are now mainstream.
But using AI in credit modelling, fraud detection, portfolio monitoring, compliance and operational risk is considerably more demanding.
These applications require high-quality data, governance, explainability and regulatory oversight.
Financial institutions cannot simply deploy opaque algorithms in areas that determine whether customers receive loans or whether suspicious transactions are blocked.
Serbia therefore faces the same challenge as many European markets: technological capability is advancing faster than institutional governance frameworks.
The opportunity is considerable.
A banking system expanding credit by double-digit percentages generates enormous quantities of customer and transaction data. Better analytics could improve early-warning systems and allow banks to identify emerging credit deterioration before conventional ratios move.
AI could also reduce compliance costs, improve anti-fraud controls and accelerate corporate credit analysis.
The objective should not be rapid adoption for its own sake.
The strongest Serbian banks will be those that combine traditional balance-sheet discipline with carefully governed technology.
The financial sector has already solved many of the capital-quality problems that defined its earlier development stage.
Its next competitive divide may increasingly be determined by data, automation and analytical capability.
