AI Tools for Accountants (#379)
/The topic of this episode is artificial intelligence tools for accountants. I could go into the background of how AI models are generated, but what’s of more interest to practicing accountants is how to actually use these models in their day-to-day accounting activities. What I’m going to do is describe some of the products that are out there right now, so you can get an idea of what’s available.
AI for Financial Planning and Analysis Software
Let’s start with financial planning and analysis software, and I’m going to use Datarails as an example. What they’ve done is layer AI on top of their software, so it runs an automated trend and predictive analysis to spot issues that you might not otherwise see. That’s useful for flagging problem areas before they escalate. For example, it might spot sales cresting for a particular product way down in just one region, which you might not see by manually scanning the data – which means that you could then take steps to scale back on the inventory for that item in that particular region. In addition, it uses AI to automatically generate presentations from the underlying data. This can keep you from spending hours creating a PowerPoint presentation for upper management. So in this case, the AI component provides value by spotting trends and anomalies, as well as by saving time in constructing presentations.
Datarails is not the only company using AI in its financial planning and analysis products. Oracle and Planful have similar capabilities, and I’m sure you can find other products. The main issue here is that these are higher-end products, where customers expect these companies to be on the leading edge of innovation. So, unless you’re willing to spend a lot of money on software fees, you probably won’t have access to these features just yet. Over time, of course, a lot of these features will probably trickle down into less expensive products.
AI for Auditing Products
Let’s switch over to auditing products. All of the Big Four audit firms have auditing software with AI features. Deloitte operates Argus software, while Price Waterhouse has Halo, EY has Helix and KPMG operates Clara. Each one of these systems has an AI engine that reviews large datasets for risk scoring, and irregular transactions, and in at least one case, even scans client contracts to see if they’re dealing with revenue recognition correctly. The great advantage here is that these systems can scan the entire population of a client’s accounting data, rather than just a sample, which massively reduces audit risk.
Operating these systems is expensive, but it’s cost effective for the Big Four, since their clients are big and rich, and so can afford the fees.
There are similar systems available for smaller audit firms that can’t afford to develop their own systems in-house, such as Inflo Audit and Mindbridge AI Auditor. All of these systems are good at flagging unusual transactions, which reduces audit risk. And, some of them even auto-generate workpapers based on your past engagements. So in these cases, the AI component provides value by spotting anomalies, and by reducing the drudgery of creating workpapers. Which is pretty much the same general types of capabilities that we just saw with the financial planning and analysis software.
AI for Accounting Software
And what about standard, off-the-shelf accounting software? AI is starting to appear down at the low end of the market, where QuickBooks Online detects anomalies in your bank feeds, and can even predict when your customers are going to pay their bills – which is kind of nice for cash forecasting.
Lets try another lower-cost package – Zoho Books. It has a smart journal entry feature that recommends which accounts to charge transactions to, which is useful for recurring entries.
Okay, let’s move up market a bit, to Microsoft Dynamics. They have an AI feature that automates bank reconciliations. And NetSuite, which uses AI for automated payment approvals, and also uses AI to assist with matching incoming customer payments to outstanding invoices.
Let’s go way up-market. Oracle’s enterprise resource planning software has an AI feature that evaluates employee expense claims for policy violations or fraud. And let’s not forget SAP, which has a bunch of high-end features. It has predictive accounting, which projects future revenue and expenses using order data and historical trends. Or, and a personal favorite of mine, it uses AI for three-way matching, where the system automatically links supplier invoices with purchase orders and receipts – which can eliminate a pile of work in the accounts payable department.
If there’s one common feature that the AI applications have in accounting software, it’s drudgery elimination – which is to say that it doesn’t necessarily eliminate tasks, but it does present you with the likely best way to proceed – such as with a proposed recurring journal entry, or a proposed three-way match. You still have to review it, but the system now takes away a good chunk of the work. Over time, this may result in some reduction in the low-end work of the accounting department, but on the other hand, you may need better-trained people who understand how the systems work, and who are mostly interested in dealing with anomalies.
And finally, I just have to bring up an earlier prediction, which was in Episode 237, back in 2017, when I took a guess at how artificial intelligence might impact the accounting profession. I’m going to rate that one a swing and a miss. I predicted that artificial intelligence would be used to automate collection calls. Nope, none of the current collection software packages do that. Instead, they use AI to recommend which collection method will work best, and the best timing for contacting customers, and for suggesting follow-up strategies. Which all represent modest efficiency gains, but not the home run of turning over the entire collections effort to a computer. Which concludes my baseball analogies.
What I’ve just brought up represents what a bunch of product designers all think are currently the most cost-effective AI enhancements that can be applied to the accounting profession right now. As time goes by, I’m sure we’ll see more advances. So stay tuned.