Unlocking the value of autonomous AI

Artificial Intelligence (AI) is putting everyday business processes under the microscope like never before. I’m not talking about headline-grabbing innovations or boardroom strategies, but the routine, run-of-the-mill workflows upon which all organisations rely.

 

In finance departments, this could mean reconciling accounts, preparing financial reports or navigating the month-end close. They’re the bread and butter of any finance team, but despite all the investment in increasingly sophisticated accounting platforms, many of these essential tasks are still based on manual processes.

 

In part, this is due to the fragmented environment in which finance teams are expected to work. Rather than operating within a single, integrated platform, many find themselves switching between enterprise resource planning (ERP) systems, spreadsheets, automation platforms and – dare I say it – emerging AI tools.

 

What’s more, this patchwork of processes is often built up over time to create something that is deeply embedded within an organisation. So when people are challenged as to why this is the case, you’ll often hear them say, ‘Well, we’ve always done it this way’ or worse, ‘If it ain’t broke, why fix it?’

 

The snag is, if organisations want to use autonomous AI to streamline their operations, they first have to address these processes. And that’s when things start to get tricky. AI won’t fix a flawed process. It will simply highlight those failings and produce them at scale.

 

 

Automation starts with better processes

In other words, while the accountancy profession is keen to adopt AI, it cannot do so without a root-and-branch assessment of its processes. And that’s easier said than done because finance teams are often so busy delivering the next close that they rarely have the time to rethink the processes underpinning it.

 

The reality is that stepping back from the day-to-day can feel like creating even more work. Taking people away from routine tasks, even temporarily, is a difficult decision when deadlines don’t stop, and workloads remain high. But without creating that space, organisations often find themselves trapped in a cycle of maintaining inefficient processes rather than improving them.

 

Which is why some of the most successful finance transformation projects I’ve witnessed began when leaders deliberately carved out dedicated time for teams to step away from their day-to-day responsibilities and focus solely on reinventing and streamlining their processes. It wasn’t easy in the short term, but it created more sustainable, scalable ways of working over the long term.

 

No emails. No meetings. No interruptions. Instead, they’re given the time and space to assess each process and decide which should be improved, removed or automated. Indeed, a 2025 report by McKinsey & Company found that “redesigning workflows was a key success factor” for the successful adoption of AI.

 

 

The changing role of accountants

Of course, one of the things that people overlook is that such changes not only impact the way an organisation is run but also fundamentally alter the role of accountants.

 

For years, finance teams have operated in what many describe as a "human-in-the-loop" model. This approach means that people are always responsible for reviewing individual transactions, checking outputs and validating results. It’s typically been a heavily manual workload.

 

Increasingly, however, organisations are moving towards a "human-on-the-loop" approach. Rather than reviewing every journal entry, reconciliation or exception manually, finance professionals oversee the wider process, stepping in only when something falls outside expected parameters or requires professional judgement.

 

By taking routine, repetitive work off their desks, autonomous AI gives finance professionals more time to focus on higher-value activities, whether that’s interpreting financial performance, supporting better business decisions or helping shape future strategy.

 

While this makes perfect sense on paper, there is some nervousness about this approach, especially regarding trust and accountability.

 

After all, in finance, AI cannot afford to be "mostly right". No auditor is going to accept "the computer told me it was right" as an explanation for inaccurate reporting, which is why governance, trust, auditability and human judgement remain fundamental to autonomous accounting.

 

Indeed, many of the latest AI platforms are designed with accountants in mind, embedding governance, auditability and human oversight from the outset to help build trust in AI-generated outputs.

 

 

Making autonomous AI work

For those finance teams ready to dip their toes in the water, the obvious question is: where do we start?

 

You won’t be surprised to hear that my advice is to start small. Pick a single workflow – perhaps account reconciliations or financial reporting – and interrogate every stage of it. Why does this approval exist? When was it introduced? What risk is it managing? If it disappeared tomorrow, what would actually happen?

The worst thing that can happen is that in the rush to embrace AI, organisations focus too much on the technology and too little on the workflows that underpin it.

 

After all, AI is remarkably good at executing the processes it is given. But it has no way of knowing whether those processes are fit for purpose. And left unchecked, organisations risk making bad processes faster rather than improving finance functions.

 

That’s why the conversation about autonomous AI shouldn’t begin with technology. It should begin with the process. That means challenging every workflow. Remove unnecessary complexity. Simplify wherever possible. Once you have the right process in place, automating it becomes the next logical step.

 

---

 

Hugh O’Neill is Principal, Accountant in Residence at FloQast

 

Main image courtesy of iStockPhoto.com and da-kuk