Hidden in the back room much of the time, subject to obscure rules that nevertheless have the potential to cast a company in an unexpectedly bad light, often relying on long and tedious manual processes - that's the image of the accounting profession. And with recent legislation making climate accounting compulsory for some companies, the complexity and risk of error has gone up again.

Can AI, seen as the solution to manual processes, help out here? People Matters asked Jairam Padmanabhan, Head of Product for IRIS Software Group, about the role AI could potentially play in accounting. He highlighted a use case that seldom comes to mind until something goes wrong: accounting discrepancies.

"AI can quickly analyse large volumes of data to identify areas of discrepancy, detecting anomalies and drawing attention to patterns that deviate from the norm," Padmanabhan explained. He laid out the most common example: businesses that have a cyclical sales cycle. AI can detect small anomalies and help pinpoint reasons for these changes. This, he said, allows accountants to help their clients recognise shifts in customer behavior, enabling better planning.

Unsurprisingly, the most manual parts of accounting are the most vulnerable to discrepancies. These are accounts payable/receivable, inventory management, payroll, and financial reporting, simply because the volume of manual data entry within those areas increases the likelihood of errors. However, Padhmanabhan said, AI can help mitigate these errors by automating data entry, flagging inconsistencies in real-time, and ensuring compliance with accounting standards.

AI's contribution: speed, efficiency, standardisation

"AI will reduce the time it currently takes to resolve discrepancies due to quickly identifying anomalies.  Previously it would have taken an accountant a lot of time to go over historical data and reports to identify those anomalies, whereas using machine learning, AI can help identify those areas quickly," he listed one of the advantages.

In addition, the speed of anomaly detection means that a sufficiently advanced tool can provide proactive and prescriptive advice on potential ramifications and allow the accountant to take appropriate action.

"For instance, AI can analyse historical data to forecast future revenue, alerting accountants if additional sales are needed to meet payroll, helping clients act proactively," he said.

And AI, with its ability to manage large quantities of data, can help with that most tricky part of the accounting cycle: the audit.

"AI can ensure that transactions and records are reviewed against the same set of historical data, reducing variability and subjectivity in manual audits. Additionally, AI can document all findings and actions, providing a clear audit trail," Padhmanabhan said.

Will AI replace accountants?

Not entirely, according to the professional associations. Even though the World Economic Forum listed accounting as one of the jobs most likely to be impacted by AI, that impact is likely to fall primarily on manual tasks, primarily basic bookkeeping ones such as data entry, ledger entry, or reconciliations. AI can already perform these accurately and up to professional standards - in 2023 Accounting Today made ChatGPT take the CPA examination, and it passed on the second try.

AI can also make analysis of historical data significantly easier, improving on the quality of business and financial planning and forecasting, and speed up the process of developing and presenting insights - one of the more important higher-level duties of an accountant.

But the human part of the job, particularly the professional judgement which accountants and specifically auditors are required to exercise, cannot be outsourced to a large language model. And few businesses, agentic AI notwithstanding, are likely to be comfortable with letting the AI call the shots when it comes to their balance sheet.

As Padmanabhan put it when we asked about the role of human expertise: "It will be up to the accountant to use their experience and knowledge of the client to utilise the information provided by AI to support the suggested action."

Ultimately, the human accountant is the one to put their name, their credibility, and their professional credentials to a set of financial statements; while AI makes it easier for them to draw up those statements, the day when the technology, not the human, signs off on an audit report is still far away.