TECHNOLOGY
AI – the future of PSA compliance?
Rob De La Rue, tax technology partner and Joe Pickering, associate director, employment taxes, RSM UK explore the helpful role of artificial intelligence (AI) in the preparation of pay as you earn settlement agreements (PSAs)
PSAs PSAs have long been a valuable tool for employers, allowing them to provide certain benefits to staff without the employees having to pay tax and National Insurance (NI) on those benefits. PSAs also remove the requirement to report minor, irregular and impracticable benefits on an annual form P11D. Instead, the employer enters into this formal arrangement with HM Revenue and Customs (HMRC), allowing the employer to settle any tax due on expenses and benefits provided to employees, with one annual submission and payment to HMRC. In addition to reducing administration for the employer (albeit at an increased employer tax cost), a PSA is an important tool in managing employer tax compliance, reducing the potential exposure to compliance failures, penalties and interest. It’s normally difficult to keep track of the tax reporting requirements of all the minor benefits provided to employees throughout the tax year, such as: l staff drinks and entertaining events l taxable travel l vouchers l small gifts l rewards l incentives. Many employers will therefore have
a PSA covering a range of potentially taxable items, allowing them to undertake a review at the end of the year to ensure all taxable items have been identified, reported to HMRC and taxed appropriately. It’s also an important tool in helping employers demonstrate to HMRC that they understand their tax obligations and have appropriate processes and controls in place to manage this. For large employers, managing the PSA process can be time consuming. This often involves manually reviewing hundreds of thousands of lines of expense data and ledgers to identify taxable items that require reporting on the PSA. This includes both those which may have been coded to a specific ledger, like staff entertaining, and those which may have been coded elsewhere. These then need to be appropriately reclassified into the agreed HMRC PSA categories, ready for reporting to HMRC. This manual review process
increases the risk of human error, including oversights, misclassification or simply restricting the data subject to review. AI AI is becoming increasingly accessible and may be the most disruptive technology for payroll and tax professionals since the introduction of the spreadsheet in 1979. It reduces a task which might have previously taken hours or days to something that can be completed in seconds. AI can be used as a ‘classifier’ of data, so it’s easy to see its application to the PSA process. For example, AI could help identify classification of expense entries containing staff drinks as a taxable item for inclusion in the PSA, even where these may be recorded incorrectly on the employer expense system (for example, under subsistence or networking, rather than staff entertaining). AI can streamline the decision-making
“Artificial intelligence is becoming increasingly accessible and may be the most disruptive technology for payroll and tax professionals since the introduction of the spreadsheet in 1979”
| Professional in Payroll, Pensions and Reward | April 2024 | Issue 99 42
Made with FlippingBook - Online magazine maker