By Jessie Taylor
The South African Revenue Service (SARS) is increasingly using artificial intelligence (AI) and machine learning as tools in its tax compliance strategy, marking a significant shift toward data-driven enforcement. These technologies have proved efficient in identifying non-compliance, and SARS’ AI-driven initiatives have helped secure billions in additional revenue.
By embracing AI, SARS has revolutionised its tax compliance framework, allowing it to track financial transactions, assess risks, and spot discrepancies with speed and precision. According to SARS Commissioner Edward Kieswetter, the integration of AI has enabled the agency to identify and prevent fraudulent activities in record time, exemplified by a recent case where AI detected R10-billion in impermissible refunds within seconds.
This powerful technology also allows SARS to access a wealth of third-party data. Every year, the tax authority gains more access to such information, reinforcing its data-driven approach. AI-powered analytics can evaluate income levels, spending patterns, and asset declarations by correlating bank statements and digital financial records, thus ensuring that SARS leaves no stone unturned in identifying undeclared wealth. This approach significantly broadens SARS’ capacity to enforce compliance far beyond the limitations of traditional audit methods.
SARS' deployment of AI and machine learning has yielded remarkable results. At the end of March 2024, SARS reported a R1.74-trillion revenue collection for the fiscal year, exceeding projections by R10 billion. The agency’s compliance programme alone contributed R293.7- billion, reflecting a 26.7% increase from the previous year. These results underscore the profound impact of AI in enhancing SARS’ tax collection efforts, as the technology identifies non-compliance swiftly and reliably.
AI tools have also enabled SARS to broaden its tax base, adding over a million individuals and tens of thousands of businesses and VAT vendors to its registers. For instance, 1,500 newly registered companies contributed R214- million, while the 39,900 new employers registered for PAYE contributed an additional R3.4-billion. These results underscore the AI-enhanced system's effectiveness as SARS strives to expand its reach and identify all liable parties within the tax ecosystem.