The SA Revenue Service (Sars) is on the hunt for cutting-edge AI technology to improve its efficiency, which it has said is at the heart of its bid to recoup about R800bn in uncollected taxes.
The tax agency in its request for information said it was looking to procure AI capabilities that would enable the utility to “replicate and automate complex, human-logic decision-making processes across its operations”.
This is as Sars said it aimed to use AI in automating an administrative decision-making process that involved the submission of reasons and supporting documents by individuals seeking an outcome.
“Once submitted, these cases are subject to evaluation, verification and decision outcome — this process is often high volume and resource intensive. By leveraging technologies such as AI, machine learning, and natural language processing, the solution aims to transform this human-intensive process into efficient, scalable, consistent and intelligent digital workflows that enhances operational efficiency.”
The agency is looking for a solution that would seamlessly integrate with its existing tax administration systems, and enable future scalability to other Sars business processes.
“Sars modernisation 3.0 requires us to build an intelligent tax and customs administration platform and broader tax ecosystem that depends on and embeds advanced data science and AI, while augmenting the work of our employees with insights from data,” it said.
The requirements by Sars might force it to look beyond the shores of SA to get a suitable service provider with domestic AI capabilities still at a nascent stage. Tax authorities around the world are exploring generative AI, though most efforts are still at an early, experimental stage.
Some countries have made progress, such as South Korea which has deployed an AI guide to help citizens file and pay taxes, while in France, AI can analyse incoming emails and propose draft responses for civil servants to validate.
Sars commissioner Edward Kieswetter earlier this year argued against a VAT increase, saying investment in improving the agency’s administration rather than increasing tax rates was a better option to improve the collection of tax revenue.
He pointed out that tax estimated at R800bn remained uncollected annually from individuals and companies. He said there was a significant opportunity to collect more taxes already due to the state instead of hiking taxes.
The National Treasury allocated R7.5bn in the medium term to Sars, with about R3bn allocated to technological upgrades, as the agency sought to use advanced data analytics and AI to detect tax-compliance risks, “close the tax gap and improve overall compliance rates”.
Sars has already begun deploying AI, an activity it is now trying to scale as demands on the fiscus increase. It used AI and data analytics to recover R89.3bn from 26-million taxpayers in the 2022/23 financial year.
In an opinion piece in Business Day earlier this month, Kieswetter said AI had revolutionised tax collection, eliminated the need for 60% of individual taxpayers to file returns, and produced assessments in under five seconds.
“AI models also detect fraud, profile risks and identify unregistered taxpayers, contributing more than R1-trillion in additional revenue over six years. Yet, these tools require human oversight to ensure fairness and ethical application, skills honed through rigorous education,” he wrote.
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.