For years, indirect tax compliance ran on a simple rhythm. Close the period. Gather the data. Reconcile. Adjust. File. The return was the finish line, and everything worked backward from it.
That rhythm is breaking down – fast – across Australia, Asia-Pacific and the Middle East. I’d been hearing it from customers for a while, and I heard it again, more sharply, when I sat down with two colleagues to record a webinar on the future of indirect tax compliance.
“The biggest shift is that indirect tax is moving closer and closer to real-time compliance,” said Vincent Lam, Indirect Tax Specialist at Thomson Reuters. Singapore’s InvoiceNow, the UAE’s e-invoicing mandate, India’s centralised clearance model – each works differently, but they share one effect. Tax authorities are also now seeing transactional data and reviewing it with AI.
“The post period operating model is now under pressure,” Vincent explains, “because governments are shifting towards real-time reporting, and that changes the frequency and the exposure of reporting.” It’s not just that returns need to be right. Every transaction behind them does too. I’m old enough to remember when a filed tax return could sit for years before anyone looked at it twice – and sometimes they never did. That world is gone.
A cross-functional problem beyond the tax team
The pressure doesn’t stay inside tax. Vincent was clear about this on the webinar: real-time reporting touches finance, IT, procurement, and shared services, because it runs through master data, order-to-cash, and procure-to-pay – not just the tax reporting process of the record-to-report chain. It has real commercial impact. If you can’t provide the invoice, you can’t collect the cash.
Meanwhile, it seems the people closest to the work are still spending most of their time on the wrong part of it. Helal Alrefai, Thomson Reuters’ indirect tax go-to-market manager, pointed to internal research showing tax professionals can spend up to 60% of their time on data preparation and administration – not on the analysis and judgment calls that actually need their expertise.
Regulators aren’t waiting for that gap to close. Australia’s ATO uses Combined Assurance Reviews to test the governance and compliance of the systems behind a return, not just the numbers on it. Singapore and the UAE have both set mandated e-invoicing timelines. The direction is consistent even where the mechanics differ.
Why the AI question is really a “where” question
Tax authorities are already using AI to find patterns across huge volumes of data – the ATO’s own AI transparency statement says as much. That cuts the other way for taxpayers too: an unusual transaction gets surfaced faster than it used to, which means the explanation for it needs to exist before anyone asks, not after.
This is where a lot of AI conversations in tax go generic – and where Helal drew a sharper line. “Tools like ChatGPT or Copilot are general-purpose AI,” he said. “They have no connection to your compliance data, your transaction history, your tax rules, your exception workflow that’s unique to your business.” Useful for drafting or summarising. Blind to the actual compliance position.
The distinction he kept coming back to wasn’t about which AI is smarter. It was about where the AI sits. “The most important difference isn’t at the filing stage,” Helal said. “It’s what happens earlier in the chain – before you even get to filing.” An AI that lives inside the workflow, with the data, the rules, and the history already there, can flag an inconsistent tax code or a missing field days before it becomes a filing problem. An AI sitting outside that workflow can only react to what someone brings to it.
Getting it right upstream, not fixing it downstream
So what does this mean in practice for teams that can’t rebuild everything overnight? Vincent’s advice, which I’ve seen play out with our own customers, was less about technology and more about sequencing: map the process before automating any of it, find where the manual, repetitive work actually sits, and prioritise the foundational issues – usually data – before any downstream initiatives.
“What this means for indirect tax teams,” Vincent said, “is that they need to get tax correct upstream – do it as close to transaction as possible. Don’t wait for it to be a period-end review. And start focusing on indirect tax compliance as part of a solution, not as the outcome.
Where this is heading
This is the thinking behind ONESOURCE Indirect Compliance powered by CoCounsel, which launched on 3 July. It’s built around the same principle the webinar kept returning to: agentic AI embedded directly in the compliance workflow, with every automated action explainable and reversible, so the professional signing off on a return can trace the reasoning behind it. Built on Fiduciary-Grade AI™, it’s designed to hold up when a tax authority – or an auditor – asks how a figure was reached.
As Helal put it: “The question is no longer whether AI has a role in indirect tax compliance. The question is whether the AI you adopt is embedded in the right workflow, with the right controls, and the right human oversight.”
That’s really the shift underway. Not less scrutiny of the numbers – more scrutiny, earlier, on data most teams used to only look at once a period. In more than 25 years across tax and finance roles in this region, I’ve watched plenty of transformation cycles come and go. The ones that stick are the ones that get the fundamentals right before the mandate forces the issue. The businesses that adjust to that now, before the next wave of mandates lands, are the ones that won’t be caught explaining themselves after the fact.
