Nicosia, Cyprus. DataArt says finance is among the most promising areas for AI adoption but also one of the most sensitive, requiring governance and human judgment alongside automation. Sergey Smirnov, Group Financial Controller at DataArt, outlined a four-step approach the company used to introduce AI in its finance function and reported measurable business results.
AI adoption in a sensitive function
Finance is increasingly presented as a response to efficiency challenges, talent shortages, and rising compliance pressure. According to Smirnov, the function can gain significant value from AI, but it also carries substantial risk because finance and accounting are data-intensive, heavily regulated, and highly sensitive to errors.
He said AI is often promoted as a tool for faster closes, improved forecasts, and automated compliance, but finance cannot rely on outputs that are only nearly correct. In this context, he said the main challenge is governance rather than technological capability.
Areas where AI delivers value
According to Smirnov, AI is most effective in finance where processes are repetitive, data-heavy, and rule-driven. He said its strongest contribution is in automating preparation work, supporting analysis, and accelerating research, while leaving final decisions to finance professionals.
At DataArt, he said AI has shown practical value in payroll and HR reporting, billing and expense verification, transfer pricing documentation support, tax research and drafting, audit preparation, and internal finance knowledge management.
Focus on business outcomes
Smirnov said the effectiveness of AI in DataArt’s finance function has been measured through business outcomes rather than experimentation. He said the company’s structured implementation produced measurable improvements, while maintaining the need for human judgment in final decisions.
