The implications of artificial intelligence (AI) in accounting is a hot topic (demonstrated by the ICAEW publication on the subject LINK) and, as with every technological advance, it offers both an opportunity and a threat to the profession.
The opportunity is to remove all the tedious tasks from the processes, freeing up resources from activities that add no value to the business and using the available resource to undertake more productive and useful work. The threat is that the released resources may not have the skill set necessary to make use of either the time or the information that is being delivered automatically through AI automation.
This is nothing new in business.
Back in the day (less than 40 years ago), every organisation had typing pools where you wrote your correspondence by hand, took it to be typed, edited the copy, had it re-typed and then signed and sent the document. It is inconceivable today for any organisation to have retained this archaic way of working. Now compare that to the way your business processes purchase invoices.
Most invoices today arrive by email as a PDF document, which are printed out on arrival and combined with the paper invoices received by post. Depending on how disciplined your processes are, the invoices are then logged into an invoice register. The paper copy is stamped, the date and general ledger codes added, before it is despatched out to the approver. The approver then signs it (when they have time) and send it back. The invoice is then keyed into the accounting software ready for payment and the paper invoice is filed away where it is difficult for anyone to access.
In this process, about 75% of all the work being undertaken (in between dealing with all of queries from internal staff and credit control calls from the suppliers) adds no value to the business at all. The only thing that has been achieved is to capture, approve and post a purchase invoice and then hide the invoice away in a filing cabinet.
AI has already fully replaced this methodology and in doing so, released at least 70% of the accounts payable function’s time, delivered a much faster and convenient electronic approval process (with simple management of queries) and posted the approved invoice to your accounting software with a far better audit trail.
The AI components include the digital capturing of all the information contained within the invoice to generate an e-invoice. The automatic selection of the general ledger coding and associated approval process for most of transactions where this is predictable (usually 80% plus of all invoices) and on final approval, the automatic posting to the accounting software. As there is now no paper, there is also no filing. The accounts payable team has not touched these invoices…at all.
Simply put, the accounts payable team is no longer involved in at least 60% of the process and can focus on the exceptions that do require human input of general ledger coding, selecting the correct approval process and dealing with queries.
Whilst this would seem to be a “slam dunk” of a commercial decision, it is what else that this approach delivers that is more important than the savings to the accounts payable process. Every approver can now see every invoice from the moment it is captured, through the approval process, when it has been posted and when it has been paid. By default, they can also see exactly what they spend with every supplier, instantly access every invoice they have received and see the exact status of their corporate budget and project information on demand.
Yet according to a recent study, 67% of UK companies have yet to make the change, effectively ensuring the “typing pool mentality” is alive and well within in the accounting function. It begs the question that if the modern financial director (FD) is unwilling to adopt AI technology for this basic process where the commercial benefits are beyond question, how long will it be before more sophisticated AI innovations are adopted?
As important, what will the consequences to the business as a result?