The Briefcase 18: Don't Panic! Preparing Your Finance Team For AI

According to Ravical’s recent report Accounting in the age of AI, 70% of firms are looking at how to embed AI into their workflows, and 68% are investing in new tools to improve productivity and client service. So, if you’re not working with AI now, the chances are that you will be soon.

As you’ve no doubt heard plenty of times already, it’s crucial to get the jump on this. You know your work best, so you’ve also got the best idea of what AI can help you with, and what it definitely can’t.

We aren’t trying to scare you – we already did that a couple of weeks’ ago with our Halloween edition – we’re just trying to help you get ready for some changes. You don’t want your clueless boss making all the decisions, so that means you need to get involved, quickly.

Step 1: Strategy before software

AI isn’t a magic wand you wave over your finance team and suddenly they’re all geniuses. It’s more like IKEA furniture – you need to assemble it yourself.

Start by asking:

• Where are your current pain points? So, stuff like invoice processing, reconciliations, and the monthly reporting cycle that makes you want to start hibernating.
• What’s your goal? Reduce manual work? Improve forecasting? Finally bypass the necessity of that one spreadsheet that crashes every time someone breathes near it?

And don’t forget to engage stakeholders early – especially IT. In fact, very rarely do you not want tech on your side.

Step 2: Tools and training

Do your research. AI tools for finance are not one-size-fits-all. You’ve got machine learning platforms, robotic process automation (RPA), and generative AI tools like ChatGPT. In fact, your existing software probably has a little bit of AI in it, even if you don’t use it yet.

And then you need to provide some training. If your team’s googling "what do when ai isnt working ahhhh”, you’ve failed them. There are plenty of modules you can find online of how AI works for accountants – there’s plenty of CPD out there covering this, if we do say so ourselves.

Part of training is accepting AI’s limitations. One of the fears of AI is that it won’t make us all more productive, it just might make the most annoying person in your office extremely productive. So, it’s worth reminding people that you can’t have AI write 400 emails a day for you without having the rest of your team despise you.

Step 3: Ethics, and not getting sued

Without a strong understanding of your ethical commitments, you’re one data breach or biased algorithm away from a PR nightmare.

• Data privacy: Anonymise data when using generative AI!
• Bias: If your data’s dodgy, your AI will be too – but faster and more confidently wrong.
• Transparency: AI can’t be a black box. People need to know how it made that forecast.
• Accountability: AI helps, but it shouldn’t be making any decisions.
• Security: Cybersecurity remains pretty important, actually!

Consider an AI Code of Conduct. No one will read it, but legal will sleep better.

Step 4: Pilot, then scale

Before you try to automate your whole finance function, start small. Think "pilot project,” not "Skynet rollout”.

Choose one area, like budget variance analysis or invoice processing. Then run a trial. Track KPIs, like time saved, error reduction, or number of sighs per task. With this data, we can learn what can be done, and what can’t be done.

Finally, scale it up like a good budget forecast – slowly, and with a backup plan.

Step 5: Watch it like a hawk

AI isn’t a set-it-and-forget-it deal, you’re going to need to keep a beady eye on it. You’re going to want to bear in mind:

• Time savings
• Error rates
• Cost savings
• User satisfaction
• Complaints (because someone will always miss the old system)

Set up regular reviews, but also celebrate the wins. If month-end doesn’t end with mass hysteria, you should have a (very relaxing) party.

🧠 Final thoughts

You might be praying for John Connor, but AI is likely to be around for a while yet. And if you get this right, you can slip ahead of the competition. But you can’t rush into an AI integration, and if you already have, take this as a reminder to review your systems, train your team, and ask the tough questions.

If all goes to plan, AI might end up eliminating the tedium in your workday, leaving you more time to play games on your phone. And that’s paradise.