Industries We Serve

How AI Is Changing Finance Departments

The real question for CFOs is not what AI can do. It is who decides what AI is allowed to touch.

Every conference keynote this year has promised that AI will transform finance "in the next few years." That's a strange way to talk about something that's already happening. In many finance departments running modern Dynamics 365, AI is matching transactions, reading invoices and chasing late payments today, on ordinary Tuesdays, while the team works on something else.

This post isn't a prediction. It's a plain look at what AI is doing inside finance teams right now, what it isn't doing yet and what a CFO should be asking before letting it anywhere near the books.

Ask any controller where the month goes and you'll hear the same answer. Reconciliation eats it. Someone pulls the bank statement, someone pulls the ledger and then a smart, expensive person spends hours lining up rows in a spreadsheet to find the handful that don't match.

That work is a perfect fit for AI, and it's where many teams feel the change first. In D365 Finance, the Account Reconciliation Agent matches transactions across accounts, flags the ones that don't line up and can clear routine exceptions in bulk within limits your team sets. Your people still own the exceptions that matter. They just don't have to find them by hand anymore.

The result isn't only speed. When the matching is done by a system that follows the same rules every time, the work stops depending on who happened to do it that month.

Accounts payable has always been a volume game. More vendors mean more invoices, which means more people typing numbers from a PDF into a screen. It's slow and it's where small errors creep in and quietly grow.

Today, AI inside D365 and Business Central reads incoming vendor invoices, pulls out the details and suggests how to code them. In Business Central, the Payables Agent picks invoices straight out of the inbox and drafts the purchase invoice for review, and it gets better with the vendors it already knows. Your AP clerk moves from data entry to review, which is a much better use of a person who understands your business.

Pair that with a document tool like ExFlow or Lasernet and invoices can flow from receipt to approval with very few hands touching them. The team doesn't shrink so much as it stops drowning.

Here's the part that should interest every CFO. Humans are great at judgment and terrible at scanning thousands of rows for the one that looks slightly wrong. AI is the opposite.

Modern finance tools can watch transactions as they happen and flag the ones that don't fit the pattern, like a vendor paid twice, a journal entry posted at an odd hour or an expense that's far outside the norm for that cost center. None of these are proof of fraud or failure. They're questions worth asking, and they reach your team while there's still time to ask them, instead of surfacing months later in an audit finding.

That shift from finding problems after the fact to catching them in the moment is one of the most valuable things AI brings to finance, and it's also one of the least talked about.

Nobody went into finance hoping to spend their afternoons chasing late payments. Yet cash tied up in receivables is cash you can't use, so someone has to do it.

AI now helps here too. Copilot in D365 can summarize a customer's payment history, point out who is trending late and draft a polite, specific follow-up email for your collections team to review and send. Instead of starting from a blank page and digging through three screens for context, your team starts with the story already told. Customers get a clearer message, and your team gets through more of the list before the day is done.

Put all of this together and you get the change CFOs care about most, which is a faster close. When reconciliations run continuously, invoices are coded as they arrive and anomalies get flagged during the month, there's far less to untangle on day three of the close.

Many finance teams still treat the close as a heroic monthly sprint with late nights and pizza. AI doesn't eliminate the close, but it spreads the work across the month so the last few days are about review and analysis rather than cleanup. That gives your team time to explain the numbers to the business, which is what they were hired to do in the first place.

It's worth being clear about the limits, because overselling AI is how companies end up disappointed.

AI won't fix bad data. If your chart of accounts is a mess, your vendor master has duplicates and three entities code the same expense three different ways, AI will simply get confused faster. Clean data and standard processes come first, and that's usually where the real project lives.

AI also won't replace judgment. It can tell you a number looks unusual, but it can't tell you whether a big customer's late payment is a warning sign or just a new person in their AP department. That call still belongs to your people.

And AI won't work on a system that can't support it. These tools are built for cloud D365. If you're running an older on-premises version or a heavily customized legacy ERP, most of what's in this post simply isn't available to you, and the gap grows every quarter.

The question CFOs should be asking isn't "Can AI do this?" It's "Who decides what AI is allowed to do?"

D365 gives you real controls here. You can set limits on what an agent can clear on its own, see every task it's working on in one place, review its work before anything posts and stop it with a click. Security roles can give AI broad read access while keeping write access tight. That's the right design, but it only protects you if someone thinks it through before switching the automation on. Rolling out AI without a clear approval model is like giving a new hire signing authority on their first morning.

Our founder, Jesper Kehlet, was on the original team that built Axapta, the system that grew into Microsoft Dynamics Finance and Operations/Supply Chain Management. We've watched this platform evolve from its first lines of code to the AI-powered ERP it is today, and we know which features will pay off quickly and which ones need groundwork first.

Hoalani Group helps finance teams at manufacturers, distributors, life sciences firms and professional services companies across the US, Canada, Europe and Africa put AI to work in D365 Finance and Business Central, with the clean data and sound controls that make it safe.

If you want to know what AI could take off your finance team's plate this quarter, let's talk. Visit https://www.hoalani.com or email us at info@hoalani.com.