Should you be using Artificial Intelligence in the office?

AI can offer huge benefits to SMEs, but it also has pitfalls. How should small to medium sized accountancy practices be using it? What are the costs and opportunities, and how might it affect your workforce and the way you work?

According to the 2020 World Economic Forum Outlook Report, the time spent on current work tasks by humans and machines will be roughly equal by 2025. While Artificial intelligence (AI) offers huge potential benefits, it is also likely to change forever the way accountancy firms work. Accountancy and financial services are predicted to be among the industries that will experience the greatest level of transformation due to machine learning, AI and automation over the next decade.

Embrace AI or see accountancy revenues fall

James Poyser, CEO of inniAccounts, believes that the future of accountancy will consist of two types of people: those who are technical A-players, and those who are technically competent, and have mastered soft skills.

He believes AI will change the skills accountants needs, and practices that don’t go embrace the change and look for alterative business opportunities will see their revenues fall.

He predicts that over the next ten years market consolidation and a new breed of accountant is likely to emerge – one that has technology at the heart of the business model. The new breed of accountant will be one who is able to manage AI but who uses their human skills to provide consultancy and strategy advice to businesses as well as doing the numbers. 

inniAccounts is a cloud-based accounting service with accountants on hand to help provide advice. It now helps thousands of self-employed people manage their accounts and tax, turns over £1million and is in the top 4% of accountancies by revenue. It recently won a Queen’s Award for Innovation for its real-time tax tool and has been recognised as a pioneer for its use of artificial intelligence at the Practice Excellence Awards. 

“We are pushing hard on developing AI,” he says. “We are trying to automate as many functions as possible. All functions which are just low-level work could be outsourced to AI to free up people and make better use of human talent. We have a strategy here in the office: For machines to do what they are best at and for people to do the work that they are best at.”

For example, he uses AI for VAT return checks, but overlays this with human checks to make sure items are being classified correctly. The programme then learns, makes suggestions and can change categories in the future.

Automation taking on repetitive work

“AI can do routine, repetitive tasks such as uploading personal expenses and categorising them. All that low-level entry level work will be automated in the future.” His prediction is that firms need to rethink their business model, because the majority of routine compliance work will in the near future be handled by technology.

“It is very likely that a technology vendor could enter the accountancy market and automate it, so accountants who rely on compliance work for revenue streams are going to see that disappear. They need to anticipate that they will have to write off that revenue stream within the next 10 years.”

The picture is not all gloomy, though. Humans can add value and divert their attention to more interesting tasks, such as strategic business advice, liaising with clients and helping their customers to build businesses. In addition to building automation into your five-year business plan, you also need to think about how you will upskill your staff so that they are able to embrace new ways of working which go beyond number-crunching.

“Your clients don’t appoint an accountant just for year-end results,” he says. “That is not where the added value will be. Instead, they are looking for soft skills – empathy and business advice. A good accountant would ask them what their goals were, what they’re trying to achieve and why, and help them achieve their goals. It’s not just about getting year end accounts filed on times.”

The challenge for smaller firms is that there is currently no single AI package that you can buy off the shelf to automate all functions and replace the low-level back office functions. While Zero and Receipt Bank do offer some of this functionality, James Poyser believes that there is still not the technology available to make AI affordable to the vast majority of practices.

Combining tech with human interaction

The solution is to combine AI with human interaction, says Professor Dr Mohammad Mahdavi, Programme Leader Data Science, AI, and Digital Business at GISMA Business School.

“One of the best collaboration models for AI and human beings is to imagine a “decision support system” role for AI, he says. The human agent is still the main decision-maker of the business but it is guided by AI-generated knowledge and insights.

“The AI-generated knowledge and insights can be directly translated to applications and services that enhance customer experience,” he says. “The most promising area of AI, which can build such intelligent services for almost any business including SMEs, is supervised learning. This branch of machine learning is related to build predictive models that can predict future events using past labelled data.”

By streamlining customer service and automating processes throughout the system, teams can focus on providing an excellent customer experience, says Dan Davies, CTO at Maintel, a cloud and managed services company.

Machine learning technology, such as Robotic Process Automation (RPA), has played a crucial role in the drive for efficiency, and RPA and machine intelligence will only grow in importance. “RPA is best for automating high-volume, repetitive, and rule-based computer interactions,” he says.

Cormac Ó Conaire is Chief Design Officer at Design Partners – a product design and innovation agency. He believes that while introducing AI is an important strategic step for businesses, it is essential to remember that your business is made up of staff, all of whom contribute to the overall success of the organisation.

“You often hear how AI can improve efficiency allowing organizations to reduce costs associated with mundane, repetitive tasks,” he says. “But instead of focusing on identifying these efficiencies and cost savings, look at the employees. Engaged employees are one of the greatest measures of sustained success. Employees are looking for purpose – people want to feel they’re making a difference. Find the areas that AI can augment your workforce and build from there. In the future human-AI collaboration will be commonplace in business.”

Accountability falls with the human

Humans will always be the gatekeepers and the providers of checks and balances, even when AI becomes a large part of business function, says Erick Galinkin, Principal Artificial Intelligence Research at Rapid7, which helps to automate functions and detect vulnerabilities in computer systems.”

“Humans who are enabled by artificial intelligences must treat those the same way they would treat a talented and eager, albeit junior co-worker,” he says. “Their guess may usually be right, but a more seasoned human is there to help catch the corner cases that they might not recognise.”

In an environment where machine learning is used alongside human judgment, the decision — and consequently, the accountability — ultimately falls to the human.

“It’s crucial that when we have a co-worker process, the machine is still only interpreting the data in the context of what it has been shown, and oftentimes that context is biased by the human decisions that have been made in the past,” he explains.

Further reading:

Marianne Curphey is an award-winning financial writer and columnist, and author of the book How Money Works. She worked as City Editor at The Guardian, deputy editor of Guardian online, and has worked for The Times, Telegraph and BBC.

Related articles