There are very few buzzwords being thrown about as much as artificial intelligence is in accountancy at the moment. But where exactly does the truth lie, between those who feel it will never grasp the nuances of accounting, and the people predicting the inexorable rise of our automated overlords?
A natural alliance
Tom Davenport, distinguished professor at Babson College, has done extensive work studying the business applications of AI, and sees a world of opportunities for accountancy. He reasons that accounting’s fundamentally structured set of activities with carefully specified tasks and outputs, lends itself to AI integration.
“When there is structure and repeatability, AI is the tool of choice. All of the leading accountancy firms are in the process of developing AI solutions, so I think that’s a pretty good indication of the potential the technology has for the profession,” Davenport explains.
A keen advocate of the increasing use of tech in accountancy, Matthew Boon of FUSE Accountants also sees huge potential, albeit with a significant delay before seismic advances take hold.
“It’s all about machine learning and developing the existing information, so the longer it goes on and the bigger data sets we have from people using digital software, the better AI will get,” he says. “That process will take some time.”
Kenny Fitzgerald partner at My Accountant Friend, is another person in the profession who is enjoying the nascent stages of the technology’s applications, but sees clear limitations.
Fitzgerald reflects: “I don’t, personally, see a time when accounts will be prepared entirely by AI, because of all the different notes and item types you have to input. If a company is giving to charity, for example, you need human input to register those nuances.”
Boon cites some significant challenges that could arise from unintended but unavoidable date biases. Illustrating his point through the medium of mortgage applications, he explains that all of the data from applications over the past 25 years will be very different in comparison to data collected from applications in recent years and in a drastically different economic climate.
The outsize proportion of the historic applications in the dataset will cause a bias towards that data, causing unfair conclusions about newer applications.
He says: “That’s a case of the data being correct, but being biased towards a whole generation of people that doesn’t fit with today.”
Davenport is also sceptical about the chances of a machine-led profession, but can, nonetheless, see the technology precipitating a reduction in the industry’s workforce.
Using the example of architecture, he cites the remaining need for highly skilled architects but a significant reduction in the demand for draftsmen, because of machines’ ability to perform that more elementary function.
“I think the same thing is going to happen with accounting,” he asserts. “It will be the entry level, transactional roles that will fall victim to a shrinking workforce.”
Cultural changes afoot
Like all advances, there will have to be a degree of cultural adaptation and upskilling among the profession to ensure that AI is used to it’s full potential.
Boon foresees a two-pronged approach to the skill acquisition necessary to embrace the technology. Technical training that offers accountants a fundamental understanding of how the software works and where it gets the information from will be vital.
Equally, he thinks that when AI gets to the stage that it’s making decisions and offering suggestions to solve issues, practitioners will need to learn how to have these new discussions with their clients, educating them about AI’s role in the process.
He says: “I think there will have to be an element of soft skills training to go along with the technical upskilling needed.”
Davenport is equally sure of the need for significant changes among the profession, especially in the case of older members who are less familiar with technical advances. In this regard, he has already spent some time at the coalface.
“I’ve worked with companies that have had large scale interventions to make their senior people more comfortable with AI and the related technologies. I think that would be a good idea for the major accounting firms,” he says.
The professor is sympathetic to a certain amount of reticence among a generation of business leaders that didn’t grow up with this type of technology, but thinks it is vital that they learn to embrace it.
Fitzgerald sees parallels between the advent of AI and the introduction of HMRC’s Making Tax Digital (MTD) initiative, in that it may speed up the decision making process for some practitioners on the verge of retirement.
“They just didn’t see the benefit of hours and hours training to use software that they may only use for another year or two,” he says.
Brave new world
So, how will AI affect the prospects of future generations? The next crop of digitally native, Gen Z accountants will, naturally, be more comfortable with the technology. However, in a profession predicted to be top-heavy, with fewer entry-level opportunities, how can these bright young things prepare themselves?
Davenport has a clear vision of the future: “If you’re in education now and you’re smart, you would be studying a joint honours in computer science and accounting. That’s pretty rare right now but I think it should be a lot more common.”
As a relatively young member of the profession, Boon has already witnessed giant technological leaps in his five years in practice. It’s this experience that leads him to believe anything is possible for AI.
He recalls: “When I started in practice five years ago, I was dealing with paper documents and excel spreadsheets. Now I don’t touch paper, and everything I work on is in the cloud. It shows how fast things can move.”