Compelling, perhaps, but how novel is the technology behind everyone’s favourite AI bot?
Why we need to talk about ChatGPT
There are a couple of things we need to talk about when it comes to ChatGPT.
Secondly, the technology might be compelling, but it’s not new. What everyone sees now is the publicly available and interactive face of technology that has been working and operating in the background for years.
And while this technology is powerful, that power is relative to the task at hand. It’s breathtaking because many of us have never had the opportunity to directly interact with artificial intelligence before. Now, with ChatGPT, Dall E and other AI and machine learning tools that anyone can access, people are suddenly and understandably overawed by their capabilities.
But let’s break ChatGPT down into what it does. It can assess a huge dataset (up to 2021) to quickly answer any question you might ask. From writing a poem to penning an analysis of global economic headwinds, it can seemingly do in a minute what would normally take you or I hours – including research and physically writing the piece. And it does so in a format that is accessible, colloquial if needed and easy to adapt.
Broken down like this, the fundamentals of ChatGPT have already been in use in the business world for years. Its impact since inception is simple: AI and machine learning take the grunt work out of many white-collar tasks we all complain about, leaving us ordinary humans less stressed and better able to deliver real value.
For many in accounting and finance, for instance, AI and machine learning are already hard at work only without fanfare.
Traditionally, accounts payable has been a time-consuming, labour-intensive and, often, a manual process. It requires the review of reams of documents, receipts, invoices, email and other data which may have been entered inaccurately in an effort to find instances of undercharging, overcharging or missed payments from suppliers and clients. This is no small task: we’re talking about dedicating time to reviewing documents by hand and hopefully making the right call at the end of a gruelling process.
But AI has been working in the background for some time already, improving the landscape for the accounts payable department. The same technology that powers ChatGPT can now automatically scour any documents (from emails to invoices and contracts, you name it) quickly and accurately, and identify instances which require a second look.
As you can imagine, this saves hours and hours of manual scanning. It gives those charged with auditing more time to report and recommend actions while enabling them to check more issues than ever before.
Wilko in the UK is one of many companies using AI for this very purpose already. The company, which sells thousands of products across more than 400 stores, has speeded up the process of reconciliation tenfold. Prior to AI, Wilko had to print off and manually check for outstanding or missing invoices, handwriting questions on the statement before emailing back a scanned document to the supplier.
Now, the retailer can drag and drop the statement into an AI platform which highlights any issues within seconds so it can be sent back to the supplier much faster. It’s reduced the process of a single statement reconciliation from up to 10 minutes to a minute. It can also check automatically for duplicate invoices, both historically and in real-time.
This is just one of many AI use cases across the globe that you are unlikely to have heard about.
Writers, creatives and more are looking over their shoulders as ChatGPT seems to be coming for their jobs. But they should rest a little easier. I predict that, much like it has in the finance sector (and the accounting function more specifically), we’ll all eventually settle into an equilibrium in which those fearing AI will adapt to it and, soon after, use it to enhance their lives.
Fear of AI will shortly be followed by acceptance, and it won’t take long before we wonder how we ever lived without it.
Chris Hutchins is the CEO of Profectus Group.
Comments powered by CComment