Start With What You Own: Why Finance Automation Should Begin Inside Your ERP

AI is moving faster than any technology finance has seen; automation is no longer a project for later. New tools promise to transform the close, and it is easy to assume progress begins with a purchase. Yet for most finance functions, the quickest and safest route to automation already sits inside the ERP they pay for. This article argues for a simple approach: map and use what you own first, then bring in AI for the work your system cannot do.

The Hidden Capability in Your ERP

Most organisations implement their ERP to get the essentials right: a reliable ledger, purchasing, sales and reporting. Once the system is live, the project team moves on, and the features parked for “phase two” often stay switched off.

Many modern ERP platforms now include bank feeds with matching rules, recurring and auto-reversing journals, approval workflows, invoice capture and scheduled reporting, and some offer AI features of their own. None of this needs a new contract, a new supplier assessment or a new integration.

For example, a team that keys its accrual reversals by hand each month may find that its ERP can reverse them automatically on the first day of the next period, with the original journal and its approval already on record.

Why the First Question Is Not Which AI Tool to Buy

A Gartner survey of 183 CFOs and senior finance leaders, published in November 2025, found that 59% were using AI in their finance function, with accounts payable automation among the most common uses. The risk is that a team buys a tool to do something its ERP could already do, then spends months integrating and validating it.

Before approving any new tool, it is worth asking a plain question: is this a gap in our technology, or a gap in how we use the technology we have? The answer is sometimes the second, and that is good news.

Mapping What Your System Already Does

A short, structured review is usually enough to find the opportunities. These could look like:

• List the manual, repetitive tasks in your close calendar and note how long each one takes. For instance, note every journal still keyed by hand from a spreadsheet.

• Compare each task with your ERP’s module list and recent release notes. Features added in upgrades are the easiest ones to miss.

• Ask your ERP partner or vendor which licensed features you are not using. It is a short conversation that many teams never have.

• Rank the opportunities by effort and risk rather than novelty. A well-configured matching rule can be worth more than an impressive pilot.

Starting With the Controls You Already Trust

Your ERP already holds the audit trail, user permissions, segregation of duties and approval history that your auditors look at each year. A rule configured inside the system sits within those controls; a tool bolted on from outside has to earn them.

For example, a bank-matching rule in the ERP leaves a record of which rule matched which transaction, who set it up and who approved the exceptions. That gives an auditor something concrete to test: the rule’s configuration, its change history and a sample of its results.

Treat every new automation as a change to your control environment. Document what it does, who approved it and how it will be reviewed, and keep that record with your other control documentation.

Where AI Adds to What You Own

Using what you own first does not mean avoiding AI; it means giving AI the jobs rules do poorly. AI earns its place where the input is messy or the judgement is pattern-based: reading unstructured documents, proposing matches for the exceptions that rules leave behind, drafting variance commentary, or flagging unusual entries for review.

For instance, a team might let its ERP match bank lines by rule, then use an AI tool to suggest matches for the remaining exceptions, with a person approving each one before it posts. The ERP stays the system of record, and the AI works at its edge rather than around it. In my own work I use a mixture of both, choosing whichever is easier to control and evidence for the task.

Encouraging a Culture of Curiosity

Tools only change how a team works when people feel free to explore them. Many people use a fraction of their system simply because nobody showed them the rest.

Here are key actions finance leaders can take:

• Give one person protected time each month to explore a feature the team has never used. Ask them to show everyone what they found.

• Recognise the people who remove manual steps, not only those who stay late to complete them.

• Invite your ERP partner to a short session on recent releases. Ask them to focus on finance features.

• Use a test environment to pilot changes before switching them on in live. This lets people experiment without any risk to the close.

Measuring the Impact of Using What You Own

Measurement keeps the effort honest and builds the case for the next step. Track a few indicators before and after each change: hours spent on the task, manual journals, open exceptions and audit queries raised on the process.

For example, a team that switches on automatic accrual reversals might count how many reversal journals are still keyed by hand each month. When that number reaches zero and stays there, the change has worked, and the team has evidence to share with its auditors.

Final Thoughts on Using What You Own

Some of the most valuable automation in finance is already licensed, already integrated and already inside your controls. Starting there is not a lack of ambition; it is the foundation that makes ambitious AI work safe to build.

Before your organisation signs for the next tool, take stock of the one you already have. Map it, switch on what fits, and save AI for the work that truly needs it.

The path to an automated finance function often begins with the system you log into every day.