AI is moving into finance workflows, data and systems will decide who benefits
Over the past few years, AI has started to show up in the tools finance teams already use, often in small, practical ways rather than large-scale transformation. Microsoft Copilot can analyse data in Excel and generate commentary or summaries, tools like ChatGPT are commonly used to sense-check numbers or draft variance explanations before they are shared, FP&A platforms have introduced predictive forecasting and scenario modelling to test assumptions more quickly, and accounts payable tools have become increasingly effective at extracting invoice data, coding transactions and flagging anomalies before they reach the ledger. In most cases, these capabilities sit alongside core finance systems rather than within them, which means they improve individual tasks without materially changing how finance workflows run end to end.
Even with that progress, the impact has remained relatively
That distance is starting to close, with AI moving closer to where finance work actually happens. Rather than sitting off to the side, it is beginning to show up within the system itself, woven into the workflows teams rely on every day. The shift is gradual rather than dramatic, which is why it can be easy to miss at first. Reconciliations no longer build up towards month end and instead continue matching quietly in the background as transactions flow through. Close processes start to surface unusual movements as they happen, often with context already attached. Cash positions update continuously, rather than being rebuilt from multiple sources. Over time, the work starts to feel different, with less effort spent gathering and stitching data together and more time spent reviewing what has already been prepared.
This is where things start to feel harder than expected. The challenge rarely comes down to a lack of AI tools, there are plenty of those already in play. It tends to come down to what the underlying systems actually allow AI to do in practice. Over time, most organisations have layered these new capabilities onto environments that were never designed to support them. Finance data still sits across multiple systems, with accounting in one place, inventory in another and CRM somewhere else again, each holding part of the picture. Pulling that together takes effort, and getting it to align consistently takes even more.
Processes tend to mirror that same fragmentation. The steps are generally understood and the flow feels familiar, though much of it relies on people knowing what to check, when to step in and how to interpret what they are seeing. AI can work within that environment to a point, summarising information, explaining movements and highlighting patterns, though it struggles to operate within it consistently. For it to do more than assist at the edges, there needs to be greater structure, with consistent data, clearly defined processes and decision points that the system can recognise and act on.
This is why the conversation is starting to shift back towards systems. ERP platforms have always sat at the centre of finance operations, though what is changing now is the level of intelligence being built directly into that core. Instead of exporting data into separate tools, AI can begin to work with live records, reports and transactions within the system itself, operating within the same permissions, following the same approval paths and respecting the controls finance teams already rely on.
That shift opens up a different kind of opportunity, where familiar processes begin to run in a different way. Reconciliations can be prepared continuously, allowing teams to focus on genuine exceptions. Close activities can be supported in real time, easing the pressure of compressed reporting windows. Cash flow views can reflect what is happening now, rather than what was last consolidated. The processes themselves remain familiar, though more of the groundwork is handled before anyone needs to step in, which is where the real change begins to take hold.
For finance leaders, this starts to introduce a different line of thinking, where the focus shifts away from which tools to trial and towards whether the organisation is ready for AI to operate within its systems at all. That readiness shows up in how data is structured across the business, how consistently processes are followed and how decisions are tracked, explained and governed, along with how finance teams see their role as more of the routine groundwork is handled within the system itself.
In practice, this plays out in different ways. Many organisations continue layering AI onto what they already have, seeing incremental gains, particularly across reporting and analysis. Others are beginning to align their systems and processes more deliberately, creating an environment where AI can operate within the flow of work. Over time, the change feels less like introducing something new and more like the system itself becoming more capable.
For most finance teams, the challenge now is making sense of what is real today and what is still taking shape. There is no shortage of claims around what AI can do, though far less clarity around what is already working inside finance systems and how it fits within the realities of day-to-day operations.
This is exactly what we will explore in an upcoming session with Oracle NetSuite, AI for finance leaders - what to use now and what's next, taking place on 1 April 2026 at 12:00 pm AEDT.
The session will look at where AI is already showing up inside finance workflows, what that looks like in practice and what organisations should be preparing for next. For those trying to connect the dots between AI tools, existing systems and how finance work is evolving, it offers a practical place to start.