For me, the great thing about working in finance is getting to make decisions that touch all aspects of the organization. Finance really places you at the center of everything, where decisions you make play a pivotal role in sustaining long-term business success.
That is, at least, the theory.
The reality is that, for teams that rely heavily on manual workflows, financial management can be not just burdensome but downright soul-sucking. When you spend most of your day manually coding invoices or processing receivables, there is little time for the creative and innovative work that really makes finance exciting.
But it doesn’t have to be this way. With a little help from AI, finance departments can automate the most toilsome parts of their work and make finance fun again. It’s almost ironic: AI automation makes finance jobs more interesting, not less.
Advertisement
This is the lesson I’ve gleaned as the leader of a finance department that has added extensive AI capabilities over the past couple of years. Allow me to explain by detailing how we’re using AI in finance, how it has supercharged the job experience of our employees and the lessons we’ve learned along the way about taking advantage of AI-powered financial processes.
Context: A Look Inside Retail Financial Operations
My company, PetLabCo, relies on a set of financial processes that are pretty standard for the retail industry. Our core workflows include:
- Receiving invoices
- Coding invoices
- Approving invoices
- Setting up payment for approved invoices
- Paying invoices
- Recording payments
- Reporting on invoice status and financials
Historically, we handled these tasks primarily on a manual basis. Someone entered each invoice into our payment system by hand, coded it to ensure we could track it properly, approved it and paid it. That work was tremendously time-consuming, so much so that basic invoice coding alone was the equivalent of a full-time job for someone on our finance team.
(In practice, our team members shared this responsibility so that no one person was stuck coding invoices all day long – but the fact that the process is so tedious that we had to make it a collective responsibility to get it done speaks to just how draining it can be.)
One Example of AI’s Impact on Finance
Starting about two years ago, we moved to an AI-based invoicing workflow. Using software from Stampli, AI tools read each invoice, code it based on our internal coding conventions and submit it for approval. Once approved, the invoices are automatically pushed into NetSuite and the final steps are set up to complete. All of this happens with minimal manual effort on the part of our team. As a bonus, our AI software can automatically generate reports that we can use to assess the finances of different departments, forecast payment needs and so on.
Thanks to these AI capabilities, we’ve turned a set of processes that used to take at least 40 hours of manual work each week into one that requires just a few hours of work by our team members.
The Need for Human Verification
Lest I imply that AI is a magic bullet with no limitations, I should note that AI hasn’t automated every aspect of our financial processes.
We still rely on humans to verify the decisions made by our AI software, which occasionally gets things wrong. The most common issue entails coding mistakes, which tend to occur when one of our vendors changes its invoice structure or format. If that happens, the software may not know how to interpret the updated invoice, and it could assign the wrong code as a result. Similarly, we sometimes run into issues when the same vendor serves multiple departments, making it difficult for our tools to know which department code to assign to the invoice.
We check for issues like these by having team members review invoice codes before we consider the invoices to be finalized. That’s the work that takes up a few hours of my employees’ time each week. But while manual invoice verification may not be anyone’s favorite task, spending two or three hours on this work every week beats having to manage and code invoices manually all day long, every day.
I’ll note as well that it took some time to get fully up and running with our AI software. Although installing Stampli took just a couple of days, the product had to learn our invoice coding procedures by tracking our invoicing practices for a couple of months before it was capable of assigning codes on its own. There’s no getting around this delay; the need to train is par for the course when you’re dealing with AI.
The good news is that once the solution had mastered our coding practices, it became capable of handling coding automatically, at any scale, on an ongoing basis.
Boosting Employee Happiness
As I mentioned, adding AI to our finance processes has dramatically boosted our team’s efficiency while also increasing our accuracy rates. Although we haven’t systematically tracked performance metrics since implementing AI, I would estimate that we are now able to process invoices about 99% faster, and that the AI-generated invoice codes are accurate 99% of the time.
But the biggest benefit is simply that our team members are now more satisfied and fulfilled in their jobs. No one goes to work in a finance department because they want to read and code invoices by hand all day long. It’s mind-numbing work, and a recipe for quick burnout.
With AI, we’re no longer subject to that challenge. Our team members can do work they actually enjoy and leave the boring bits to AI. It’s a win for our business and employees alike.
George Koukoumpanis is the Financial Controller at PetLabCo. He has nearly 20 years’ experience in financial management and analysis.