Read anything related to technology today, and one term crops up more than any other: Artificial Intelligence (AI). It is hailed as the ultimate game-changer for businesses. According to PwC, 63% of global CEOs believe AI will have a greater impact than the internet revolution of the last 20 years.
But what exactly is AI, and – with McKinsey and Co calculating that 86% of an accountant’s activities have the potential to be automated – how could the AI revolution impact accountancy and practising professionals in the UK?
What is artificial intelligence?
Put simply, AI refers to any computer system that demonstrably undertakes intelligent behaviour. ‘Static’ computer systems, such as databases and written documents, do nothing more than store and retrieve data. AI is a step beyond that, although a precise definition of ‘intelligence’ in this context is still disputed by computer scientists
According to the Institute of Chartered Accountants in England and Wales (ICAEW), AI technology has the potential to help accountants in three broad ways:
- Providing better and cheaper data to support decision making
- Generating new insights from the early analysis of data
- Freeing up time to focus on high-level tasks such as strategy, decision making and problem solving
- One immediate application of AI is to eliminate time-heavy, tedious tasks; particularly the rote functions so prevalent in accountancy. When human errors can be vastly reduced and the time taken to complete tasks significantly reduced, it is easy to see the attraction of AI.
In fact, some firms have already reduced or even eliminated accounts payable and receivable employees. AI processes are handling repetitive tasks such as data entry, matching purchase orders, processing expenses claims and initiating invoice payments. Eventually, payroll, auditing and tax remittance could also be handled by AI.
However, many foresee that the real advances for accountancy will come with the rise of machine learning.
What’s the difference between AI and machine learning?
Machine learning (ML) is a branch of AI that deals with algorithms which learn by themselves using data. Arthur Samuel, who first coined the term ‘machine learning’, defines it as giving “computers the ability to learn without having to be explicitly programmed.” This is a significant step forward from ‘expert systems’ – an older type of AI based on laboriously encoding the decision-making processes of human experts as complicated sets of rules. These were time-consuming to produce and brittle, in the sense of not being able to adapt well to changes in the environment.
By contrast, in machine learning, the algorithms learn how to perform tasks for themselves when they’re shown “training” data to learn from. This means that it’s possible – where the right data is available, and with the right skills – to produce ML systems that can very quickly solve a given task.
Since these models are not merely encoding human decision-making processes – instead finding their own solution – they can, in many cases, perform at super-human levels. The systems can also adapt more readily to changes by showing them new data. But it is important to note that any given machine learning system can only perform that one task it was specifically designed for. We currently have no idea how to produce AI that has broad intelligence in the way that humans do, and which can solve a broad range of tasks.
For this reason, it is highly doubtful if ML could ever replace accountants or auditors completely, but roles and workflows are likely to change radically. Auditors may spend much less time performing audits, and more time designing, reviewing, and verifying how information flows between systems. Accountants could see a similar change. Everyday data-entry tasks will become much easier, freeing up time for accountants to focus on analysis and insights.
In fact, for some accountants – even those operating in small or micro-sized firms – ML might already be part of their professional process. “Machine learning is increasingly becoming integrated into business and accounting software,” says the ICAEW. “As a result, many accountants will encounter machine learning without realising it.”
There are some real-world impacts of this to consider, especially on the training and skills needed by the next generation of accountants. As well as their financial training, accountants will need to be involved in training or testing models, or auditing algorithms, according to the ICAEW.
Depending on an accountant’s role, they may need anything from a superficial knowledge of ML through to a deep understanding of modelling and algorithms. But what is clear is that having a solid understanding of the terminology and concepts associated with AI and ML will be key to navigating both the industry hype around these developments and the changes they will bring to a typical accountancy practice. Accountants and firms that develop these skills now will be able to differentiate themselves as the technology becomes more widespread.