Finance teams are using AI more and more, whether it’s to speed things up, spot patterns or make better decisions. Early efforts often ran into the same walls: fragmented data scattered across systems and AI outputs that couldn’t be fully explained or traced. Increasingly, though, the question has now shifted from whether to use AI to how to roll it out properly.

Many CFOs are still cautious and for good reason: the things that make AI useful, such as its speed and ability to spot things humans might miss, are the same things that make it hard to trust with financial numbers that boards, auditors and regulators need to rely on.

As Robert Crawford, BlackLine’s senior director of agentic AI and international growth, puts it: in finance, every decision must be explainable and defensible. “If AI decisions can’t be traced back to accurate data, rules and approvals, risk displaces efficiency,” he argues.

Agentic financial operations has emerged as a response to that pressure by pairing AI with trusted financial data and controls built into the workflow itself. Part of the problem it solves is structural: off-the-shelf AI isn’t built with accounting logic in mind, and the data it needs is usually spread across ledgers, platforms and spreadsheets.

BlackLine predicts that by the end of 2026, businesses still relying on these kinds of disconnected systems will see 40% more corrections demanded by auditors – a clear sign of how costly messy, scattered data and processes can become.

What it looks like day to day

Three core pillars underpin the agentic financial operations model. First, a unified data layer creates a single source of truth across the finance function. Second, the agentic intelligence layer and third, an auditable system of record. This is imperative because the team that processes a transaction is never the same one that approves it. “This is the financial jurisprudence that CFOs expect and demand,” says Crawford.

Reconciliation is where this is most visible. Instead of a human working through accounts line by line, an AI agent can run the matching by pulling data from multiple systems, spotting discrepancies and packaging the results for a reviewer to sign off.

BlackLine says teams have seen time savings of up to 90%.

The transparent approach

Crawford calls this a “glass box” approach. “Without complete transparency into how an AI arrives at its conclusions, CFOs cannot confidently defend its outputs,” he says.

But most businesses aren’t there yet. Deloitte’s 2026 AI research found that while 74% of companies expect to be using agentic AI within two years, only 21% currently have proper governance in place to manage it safely.

Even so, BlackLine expects more than 70% of finance leaders across Australia and New Zealand to be ready to make this shift in 2026, positioning agentic financial operations as the model set to define the next era of finance teams.

Find out more

BlackLine is a CA ANZ Member Benefits partner. For more information, click here.

To learn more about BlackLine’s agentic financial operations, click here.