AI could be your junior colleague By AAT CommentArtificial intelligence 27 Sep 2023 AI and accountancy could be a powerful mix. Are you willing to change the way you do things? Other articles in this series Work with AI without risking it all Finding an ethical balance in AI AI technology will develop at a rate we have never imagined. Because of this, sitting on the fence is unlikely to be a fruitful strategy. According to Mark McDonald, senior director at technology research and advisory firm Gartner, things are going to change more in the next 10 years than they have at any other time in the history of accounting and finance. James Osborn, chief digital officer at KPMG UK, bills generative AI – deep learning models like ChatGPT that can generate original and realistic text, images, or other media from prompts – as “transformative technology”. 400% increase in knowledge worker’s productivity by 2030 Prediction by Ark Invest Big Ideas “I think it will infuse everything we do over the course of the coming years.” Osborn said. “The winners in this will be those with the richest internal data sets and the most employees who are willing to change the way they work.” But are we willing to change the way we do things? AI could be your sidekick Practitioners view AI as a sidekick to humans rather than a replacement (or even threat to humanity). “We see it applying to all our audit, tax, advisory and our internal uses pretty extensively,” Osborn said. “The practitioner becomes the co-pilot with the generative AI that is on your data, and they can prune and check and ensure that the model is good,” Osborn said. It’s hard to deny the thrill of seeing AI at work: instead of being faced with a blank sheet of paper, watching an AI mobile app spit out an article, email or report in under a second knowing it would have taken you many hours is definitely an eye-opener. “It’s great for a first draft. It’s like having an army of quite good people working for you 24/7. Imagine the leverage that can give a smaller organisation,” said one AI expert. Generative AI can also be a powerful aid in general research or generating creative brainstorming options. Even in the form of a mobile app for iOS or Android, AI can immediately boost personal productivity by at least 20%. It’s as easy as using Google and a lot more powerful, so it’s incredibly tempting for individual workers to start experimenting with the AI, with or without their employers’ encouragement. Perhaps employers are well advised to see what the technology can do and how to develop safe practices. How staff see the future of the workplace “If you plan on being in accountancy for more than five years, you have to change how you do things,” said Stuart Hurst, managing director at Accounts and Legal. “Staff are interested in this kind of tech. So if you want to hire the next generation of staff, you have to be moving into this field. And the final argument is to make more money! Fees are being squeezed, and wages are rising, but AI can improve margins and open up new services.” Almost nine in ten accountants believe that within the next five years, AI will play a crucial role in their ability to be an effective strategic business partner to their clients, according to research published in May by Intuit QuickBooks. And 95% of accountings intend to invest in new technologies in 2023, to the tune of an average £10,000 – £25,000 – with AI the most likely area of investment, according to Intuit. “It’s [potential is] definitely not over-blown, and we’re already using AI,” says Marc Bena, digital audit leader at PwC. “We’ve got technology called cash.AI where we audit cash using AI to read documents. That tool improves the quality of reading documents, translating end statements, reconciliations and doing the interim process of audit of cash.” However, Bena is at pains to point out that “AI here is not just making judgements, it’s purely helping improve the quality of how we read documents through Natural Language Processing (NLP).” Onboarding AI safely The main pathway for mid-sized accounting firms or finance teams to get started with AI is through off-the-shelf solutions. The advice from experts is to begin experimenting with low-risk, high-impact areas such as content generation or marketing, where there are no data privacy issues. It’s also wise to get employees to work in peer groups so they can swap learnings, spot risks and establish sensible boundaries. Big-name vendors in accountancy like Sage, Xero and Avalara are using various AI applications in their existing products. They are also beginning to tentatively enter the generative AI arena. “As AI evolves it is going to be accessible to everyone,” said Stuart Miller, head of product compliance and industry engagement at Xero. Sage recently launched its first generative AI capability – a tool modelled after an email inbox that automates workflows, assigns tasks and understands the intent of inbound emails sent to the accounting team. Alongside basic software, Rob Hackney, tax manager at DSG Chartered Accountants, says he is “seeing the beginnings” of smaller firms using complex AI. “Primarily at this stage it is the quantitative elements of it rather than going into what is more complex areas where there is advice or more qualitative conclusions coming out of it,” Hackney said. How large companies are doing it The larger accounting firms have started experimenting with the use of generative AI. PwC and KPMG have both announced plans to invest heavily in the coming years, working with Microsoft and others to automate aspects of their tax, audit, and consulting services. For instance, KPMG has partnered with Blue J to launch an AI tax analysis tool in the UK that will enable KPMG’s tax team to accurately predict tax scenario outcomes and reduce the time spent searching for and analysing tax legislation and case law. Similarly, by infusing data analytics and AI into the audit process, auditors will be able to focus on higher-risk areas of audit, analyse transactions on a more granular level and perform audits on a more real-time basis. PwC has also been forging strategic partnerships with Silicon Valley firms that are pioneering the development of generative AI, such as Harvey, an AI startup built on OpenAI and ChatGPT technology. The platform uses natural language processing, machine learning and data analytics to generate insights on various aspects of legal work, including contract analysis, regulatory compliance, claims management, due diligence, and advisory and consulting services. Hassane Ferdaous, digital audit partner, at PwC, said the firm’s work with Harvey is “bridging the gap” between large language models that are available to the public and those that can be leveraged by business. “We see generative AI as a significant augmentation of the capabilities of our people and adding value to our clients and our markets,” Ferdaous said. “The knowledge base that we have about audit and tax can augment the AI models. And that is more powerful.” Develop the skills to partner AI Mark McDonald of Gartner believes in the new AI-driven environment accountants will be expected to analyse information and understand what is happening, why it happened and what is going to happen in the future. “That is not something that arithmetic helps us with. That is something the field of data science is better suited for,” McDonald said. “The financial professional of the future will need to have skills that combine accounting and finance knowledge with statistics and programming.” KPMG’s Osborn believes accountants will need to develop the art of ‘prompt engineering’ – feeding the right information and questions into the model in order to get meaningful or accurate results. Your critical thinking will be necessary Critical thinking is also key. Adam Williamson, AAT’s head of professional standards, said one of the problems with generative AI is the risk that “if you ask slightly the wrong question and you get an answer that looks like it could be realistic but is wrong”. “That is where you need assurance. You still must do that critical thinking. The actual using of the tool is not the issue, it is the outcome, it is what you do with the information that you are being given,” Williamson said. Indeed, while AI may take on boring, repetitive, and tedious tasks, it will not replace the human factor of relationship development, and the accountant’s role as a trusted business adviser. Viewing AI as a junior colleague may be helpful. It needs clear instructions and any work it produces needs to be checked as it can misunderstand, get things wrong and occasionally even make things up. The best thing accountants can do is be tech savvy about how AI is going to be used in the profession, Vsu Subramanian, who heads up Avalara’s AI initiatives, said. “It is good to have an AI 101 and demystify what AI is. Once you do that it will start to make sense, you will be a more informed and more valuable accountant to the future of the profession because it is going to transform and you want to be part of the transformation,” he said. Human judgement Most tools like DataSnipper will have embedded AI engines that learn patterns and react accordingly. However, it’s not just about knowing which buttons to press: “All the way through the process you have to use your judgement,” says James Hadfield, head of audit at Top 30 firm Menzies. He’s an enthusiastic adopter of a tool called Inflo, which automates a range of audit and accounting tasks. This includes journal testing, the basic audit activity designed to detect fraud and where management has overridden controls. Hadfield says even in that example, for accountants using AI tools, there can’t be any substitute for good judgment and informed risk management. “When you’re setting the risk parameters of that you need to be thinking about, what is an unusual entry for this particular company?” he says. “It might be something that’s been entered by a particular person, or posted on a weekend, or anything anomalous. So even with tech it’s not as simple as just clicking buttons – you still need to know what to look for and how to use the tech in the most effective way.” AI’s CV How AI can support finance professionals. Automating routine tasks: AI can manage repetitive tasks like data entry, invoicing, payroll, tax preparation and reconciliations, thus freeing up accountants to take on more advisory roles. Risk and fraud detection: Machine learning algorithms can analyse vast amounts of data for unusual patterns or discrepancies, enabling early detection of fraud or other financial risks. Predictive analysis: AI platforms can analyse historical data and provide predictive insights on trends, cash flow forecasts and other relevant metrics that facilitate decision-making. Real-time reporting: AI allows accountants to generate real-time reports, enabling organisations to make data-driven decisions quickly. Regulatory compliance: AI can understand and be updated regarding the dynamic landscape of financial regulations. Taxation: AI systems can provide a resource for tax regulation interpretation and guidance. They can help accountants analyse complex tax laws and create a detailed tax strategy. Accountants’ future skillset How accountants will adapt to thrive alongside AI. Technological Skills: This includes understanding the basics of programming, big data management, and AI algorithms to comprehend how technology can influence and form their work. Data Analysis: AI can be used to process a large amount of data to extract useful insights. Therefore, data analysis skills can help accountants understand and interpret data obtained from AI tools more effectively. Critical Thinking: Human judgment is still an indispensable part of the job. Accountants will need to interpret and make strategic decisions based on the information provided by AI. Adaptability: Being adaptable and open to changes will help accountants stay up to date with new technologies and their applications. Understanding of AI Ethics: Accountants must understand the ethical implications of using AI, such as data privacy and security concerns. Problem-solving: AI can provide solutions, but the ability to define problems clearly and apply the AI tools effectively still rests with the accountant. Continuous Learning: AI in the field of accounting is still developing, requiring accountants to pursue continuous learning to keep up with new updates and tools. AAT Comment offers news and opinion on the world of business and finance from the Association of Accounting Technicians.