Recently I’ve been spending time looking at the various initiatives that the accounting software vendors have in terms of applying automation (via artificial intelligence) to their core offerings. While my assessment is that there is little happening in the enterprise space around this topic, in the small- and mid-sized accounting space there is huge amounts of work being done on this. I have seen development from a number of the main players in the space aligned to this automation theme and recently spent some time at Xero’s global HQ talking to their own machine learning team.
It is hardly surprising that Xero is at the forefront of talking about machine learning — its CEO, Rod Drury, has been waxing poetic for some time about the company’s move to Amazon Web Services and what that move enables in terms of utilizing the higher-level machine learning services that the Amazon cloud offers.
But beyond cool functionality being talked about, I’ve been thinking about what the end result of all this automation will be for the practitioners of note today — those accountants and bookkeepers who are responsible for the millions and millions of small- and mid-sized businesses’ and their accounts. The vendors and the accountants who see what is happening keep talking about "unlocking accountants to become truly strategic partners to the business" or words similar. but what does this mean? And will accountants really be able to do something better than a machine can?
Of course, there is some context here: in recent times there has been increasing disquietude about automation and what it will mean for workers. And while this is not new (the invention of the printing press, after all, was a threat to both the scribes who disseminated written knowledge but, more importantly, those who held power through the concentration of knowledge among a handful of elites), it is fair to say that technology is accelerating this process. One only needs to look at Uber or Airbnb to see what disruption looks like. But sadly, accountants seem to be in two camps; firstly those who don't see the change that is coming, and secondly those who somehow think disruption will magically slide off them and they'll remain unscathed.
And whereas the last few decades have been about automation threatening the livelihoods of blue collar workers (thing gas lamp lighters, ice delivery personnel, and blacksmiths) today it is the so-called knowledge economy or those white collar workers who felt secure, who are being threatened.
So if we look at what is happening in the accounting space, we can see examples of IBM suggesting that computers will replace human tax agents, Sage suggesting that its new chat bot will, over time, change the role of the adviser, and Xero’s initial forays which are simple and discreet (in its case automatically suggesting account codes on new transactions), these seemingly innocuous first steps are the tip of the iceberg in terms of sector disruption.
Recently I was talking to someone employed within the accounting software industry. This person was a highly skilled artificial intelligence practitioner and our conversation was around the applicability of these early initial forays into accounting software. I questioned this person about the longer-term impacts and whether this marked an end to the accounting industry as we know it. Their response indicated the sensitivity with which these vendors view their position, they tried to vacillate suggestion that all of this intelligence will simply mean that accountants can give better advice because “businesses always want a human interface for their information.”
I’m not convinced that this perspective is legitimate — after all, most of us have accepted that eBay and Amazon’s AI-generated suggestions are probably as good, if not better than the sort of recommendations we get from our local merchant or bookseller. And while autonomous vehicles have a degree of perception bias around them, most people, when thinking through a lens of rationality, admit that an autonomous vehicle is, in most cases, probably safer than a human controlled one.
And so the short to medium term will consist of a deepening of the breadth of what AI can do within accounting software — today it’s simply about chatbots, limited automation and smart data entry replacing more manual processes. But extrapolate that forwards a few years and we can see that there is little difference between that and, for example, generating automated business advice based on classification and analysis of anonymized mass data.
As these accounting software vendors gain more scale, and that scale happens in the cloud, they are better positioned to derive these insights. Xero, as an example, recently hit the one million customer mark and, of course, all of Xero’s customer data is in the cloud. The cost and complexity of actually generating personal insights, derived from aggregate data, with a data set of that size, is minimal. Perhaps more importantly, the quality of the advice that can be generated when you’re talking about a seven-figure customer base is very high.
And therein lies the real issue for accountants, while it may be challenging to do so, my assertion, borne out from what I have witnessed over the years, is that insights generated by computer over a million-plus customer base will be far superior to those more manually generated ones from a single accountant or practice and its lengthy history of experience. Humans are fallible, have biases and forget things. Machines never do.
Most vendors, when challenged with this assertion that the end is nigh for accountants, like to fall back on the comfortable position — that accountants and practices will simply pivot and move more into the “creative” parts of the business. They will replace transactional work with deeper insight-driven work. Synthesizing, analyzing and extrapolating information and trends for their customers.
I’m not convinced. Machines are getting better by the day. A good proxy for the speed with which accounting software will be able to provide a qualitative improvement in even the most “creative” aspects of an accountant's role lies in the AI-driven, voice-activated offerings such as Google Now, Apple’s Siri or Amazon Alexa. From their initial introduction only a few years ago when they were, at best, a novel and entertaining concept hardly applicable to the real world and general public, they are today becoming deeply embedded in the lives of even late-adopters. Microsoft, IBM and Google have all trumpeted in recent weeks about the fact that their voice recognition offerings have, for the first time, beaten the quality that a professional human dictator can achieve.
Extrapolate this speed and depth of adoption into the accounting software world, and you have an industry that is changing: rapidly and forever.