Automate and Systemise Routine Finance Work

The real value of automation isn't speed. It's giving finance more time to think.

The work that quietly consumes finance



Most repetitive finance processes don't look particularly alarming when you examine them individually.

A reconciliation might take an hour. Updating a report might take another. Preparing the data for a management pack might take half a day.

None of those activities seems like a major problem.

But add them together across a finance team, then repeat them twelve times a year, and something interesting happens. A significant amount of finance capacity disappears into work that follows almost exactly the same process every month.

That's where I think the automation conversation should start.

Not with:

What can this technology do?

But with:

What are we doing repeatedly that doesn't really require us to think?

Those are very different questions.



Start with the boring stuff



Stop processing everything



This is where I think automation starts to change the role of finance in a more interesting way.

Instead of processing every transaction, we begin managing the exceptions.

If the vast majority of transactions meet an agreed set of rules, perhaps finance doesn't need to spend its time looking at every single one.

The system deals with the routine transactions. Finance looks at the ones that don't fit.

Why didn't this supplier transaction match?

Why is this accrual significantly different from its normal pattern?

Why has this balance moved outside the expected range?

Why is this journal unusual?

Now we're using people for something people are particularly good at.

Judgement.

The aim isn't to remove finance professionals from the process. It's to stop using their time on work that doesn't need their experience.



But automation can also make bad processes faster



There is an obvious catch.

You need to trust the data.

If an automated process relies on information coming from an ERP system, invoices, payments or another source, then the quality of that information becomes even more important.

Automation doesn't magically fix poor data.

In fact, it can do the opposite.

It can allow bad data to move through a process much more efficiently.

That is why I don't think we should separate conversations about automation from conversations about data quality and controls. If you're going to remove manual intervention, you need to understand what you're removing and what controls still need to remain.

A manager may still need to approve a payment. Someone may still need to review an unusual transaction. Material exceptions may still need investigation.

We're not trying to remove judgement.

We're trying to make sure judgement is being used where it adds value.



Then Anthony described a process we've all seen



The conversation became particularly interesting when Anthony moved beyond transactional processing and started talking about reporting.

The process he described will be familiar to many finance teams.

Someone extracts information from the ERP system.

It goes into Excel.

One worksheet cleans it.

Another transforms it.

Another contains a pivot table.

Another combines the results with prior month information.

Eventually, after enough copying, pasting, checking and refreshing, the team has something that can be analysed.

Then next month arrives.

And they do it again.

We've lived with processes like this for so long that they've started to feel normal.

But there's a question worth asking.

How much of that process actually requires a finance professional?



Probably very little.

The analysis might.

The interpretation certainly does.

The preparation often doesn't.

Four or five days before anyone could start thinking

Anthony shared an example that brought this into focus for me.

One corporate client was taking approximately four to five days to move information from its ERP system to the point where finance could begin looking at the reports.

Think about that for a moment.

Four or five days.

Not to analyse the information.

Not to understand why performance had changed.

Not to discuss what management should do next.

Just to get the information ready.

After implementing a Corporate Performance Management solution, that process came down to approximately two or three hours.

It's tempting to look at that and say the organisation saved four days.

It did.

But I think that's the less interesting part of the story.

The more interesting question is:

What did finance now have the capacity to do with those four days?

That's where the real return begins.



Preparation and analysis are not the same thing



I've spent much of my career working with finance teams, reporting systems, business intelligence and analytics, and one pattern keeps appearing.

We confuse producing information with analysing it.

They aren't the same thing.

A finance team can spend days producing a beautifully formatted management pack without spending very much time understanding what the numbers are actually saying.

The report gets produced.

The deadline gets met.

Everyone is exhausted.

Then someone in the management meeting asks:

Why did this happen?

And suddenly we're doing the analysis.

That is the paradox of finance automation.

The work that looks simple to automate often doesn't look particularly valuable when viewed one task at a time. Yet removing enough of that work can create something enormously valuable.

Time to think.



AI makes this conversation even more interesting



Variance commentary is a good example.

Every month, finance teams spend hours writing commentary, updating presentations and preparing slightly different versions of essentially the same information for different audiences.

Some of that requires genuine interpretation.

Some of it doesn't.

With reliable data, clear rules and well designed prompts, AI can increasingly help with the repetitive parts. It can identify movements, draft variance commentary, surface potential anomalies and provide a starting point for analysis.

But I wouldn't hand the entire process over to AI.

That's missing the point.

The value isn't that AI can write the commentary so the finance professional doesn't have to think.

The value is that AI can help with the preparation so the finance professional has more time to think.



There's an important difference.

You probably already have more automation technology than you think



Another assumption I often hear is that meaningful automation requires a major technology investment.

Sometimes it does.

Large organisations may need sophisticated data integration, workflow or Corporate Performance Management platforms.

But that's not always where you need to begin.

Many finance teams already have Excel, Power Automate, AI and other tools capable of removing a surprising amount of repetitive work.

The first step isn't necessarily buying something.

It's understanding the process.

What happens every month?

Which steps are repeated?

Which steps follow rules?

Where are people moving data from one place to another?

Where is judgement actually required?

Once you can see the process clearly, the automation opportunities often become much easier to spot.



So, what should you automate first?



I wouldn't begin by trying to automate the finance function.

Pick one process.

Preferably one everyone dislikes doing.

Document what actually happens, not what the procedure manual says happens. Understand where the data comes from, what rules are being applied and which decisions genuinely require human judgement.

Then separate the two.

Automate the predictable work.

Keep people focused on the exceptions.

Measure what happens.

If a process that took six hours now takes one, don't stop at celebrating the five hours you've saved.

Ask what you're going to do with them.

Because capacity only becomes valuable when you use it differently.



The question I'd ask your finance team



At your next month end, watch where the time goes.

Not just whether the work gets completed. Look at what your most experienced people actually spend their day doing.



How much of their time requires judgement?

How much requires an understanding of the business?

How much involves interpreting information and helping someone make a better decision?

And how much of it involves moving data, matching transactions, refreshing spreadsheets and repeating something they did last month?

That gap is your automation opportunity.

The purpose of automation isn't simply to make finance faster.

And it certainly isn't to remove people from finance.

It's to remove work that doesn't need people, so finance professionals can spend more of their time doing the work that does.

The real return on automation isn't speed. It's capacity.

And what finance chooses to do with that capacity is where things become interesting.

Frequently Asked Questions

What finance processes should be automated first?

Start with repetitive, low risk and high-volume activities that follow predictable rules. Examples include creditor reconciliations, recurring journals, prepaid expenses, accruals, balance sheet reconciliations and recurring reporting activities.

How can automation improve month end reporting?

Automation can reduce the manual work involved in extracting, transforming, checking and preparing financial data. This gives finance teams more time to analyse performance, investigate variances and provide useful insights to the business.

Why is data quality important for finance automation?

Automated processes depend on reliable source data. Poor quality data can cause errors to move through a process more efficiently, so data quality, controls and automation need to be considered together.

Can AI automate variance commentary?

AI can help identify movements, generate draft variance commentary and provide a starting point for analysis when it has access to reliable financial data and clear prompts. Finance professionals should still review the output and apply business context and judgement.

What is the real benefit of finance automation?

The biggest benefit is not simply completing processes faster. Automation creates capacity by reducing repetitive work, allowing finance professionals to spend more time on analysis, judgement, decision support and business partnering.

Listen to Episode 14 of the FP&Ai Podcast

In Episode 14 of the FP&Ai Podcast, I explore this subject with Anthony Wilson and Donovan Moses, including practical examples of automation, exception management, reporting and the growing role of AI in finance.

If your team is spending too much of every month preparing information and not enough time understanding it, the conversation is worth listening to.