How AI is shaping the accounting industry
Artificial intelligence (AI) has become the hottest topic in tech over the past year. But what is its impact on the accounting industry and how can practices leverage its capabilities?
AI: revolution or evolution?
If there’s anything we’ve learned over the past few years, it’s that accountancy is constantly evolving. From changing legislation to technological advancements, there are always new ideas to debate and assimilate. The latest hot topic of conversation is the impact of artificial intelligence (AI) on the accounting industry.
Since the launch of ChatGPT in late 2022, there’s been a surge in interest in AI and how it will transform the working world. Rapid technological developments have promised endless benefits for accountants but have also led to uncertainty about how the profession can harness the power of these new tools.
Many employees across the global business community remain wary about AI and its potential impact on their jobs; others have concerns about data accuracy and how clients and their customers will respond to the technology.
A brief history of AI
Before taking a closer look at the impact AI might have on accountancy, let’s take a step back. It’s easy to think of AI as a new generation of chatbots and content creation tools set to drastically transform the working world. But, in reality, researchers have been developing and evolving algorithms and statistical methods for several decades.
Since the 1980s, businesses have been utilising AI to provide automated advice and decisions for complex problems using knowledge and rules set by humans. The 1990s and 2000s saw the rise of machine learning, which gained prominence in areas like portfolio management, pricing and trading strategies. Finally, the proliferation of ‘big data’ and deep-learning techniques in the 2010s are now used to power everything from fraud detection and credit scoring to translation and image analysis tools.
AI today
Today, everyday devices, from digital voice assistants to the recommendations you see in your favourite streaming service, are driven by AI. These technologies observe and learn from patterns in data and use these patterns to make predictions and classifications. Until recently, AI was limited to these predictive models.
However, recent technological advancements have allowed models to go further. Machine learning can now use data for more complex forms of prediction that create entirely new outputs through generative AI (GenAI). Generative AI tools, such as ChatGPT, Google Bard and other bots like them, are examples of large language models, or LLMs. These LLMs are trained to process vast amounts of text, spot patterns in how words and phrases relate, and then predict what should come next. Within seconds of being prompted, these tools are now advanced enough to produce unique content, including blogs, social media captions, holiday itineraries, poems, Excel formulas, images and more.
Vast amounts of data power these LLMs, and their capabilities are expanding quickly. ChatGPT-3.5 uses around 175 billion parameters, while ChatGPT-4 (currently only available as a paid subscription) is rumoured to have over 1 trillion, a significant increase that is only likely to produce more sophisticated results over time.
Don’t know your AI from your LLM? Here’s our cheat sheet.
Artificial intelligence (AI): refers to the development of computer systems or machines that can perform tasks typically requiring human intelligence.
Generative AI (GenAI): a subset of artificial intelligence focused on creating or generating new content, such as text, images or other data.
Machine learning: a subset of artificial intelligence that focuses on developing algorithms and models that enable computers to learn from data without being explicitly programmed.
Large language models (LLMs): advanced machine learning models designed to process and generate human-like text based on extensive training on vast amounts of textual data.
Big data: refers to large and complex datasets that are too large, varied and fast-moving for traditional data processing techniques to handle effectively.
How will AI help accountants?
The future of AI in accounting is projected to improve ways of working, increasing productivity and saving time on tasks you most likely already wished a machine could do for you. The most beneficial examples include tasks involving long hours, like repetitive work or ‘big data’ activities.
Automation of data processing is becoming commonplace within many accounting apps, and some systems, including FreeAgent, already use machine learning algorithms to make banking and expense reconciliation quicker and easier. Therefore, accountants should view the introduction of AI as an evolution of the tools already in place rather than as a replacement for the experts currently performing the task. As we’ll see, the human element will be more valued than ever once AI is established within the sector.
Saving time
Just like spreadsheets changed the industry in the 1980s, AI is set to transform some of accounting’s most time-consuming tasks. The concept of automating repetitive processes by using digital tools and systems is already widespread, but AI is set to offer even further improvements.
The biggest of these improvements is likely to sit within the processing of client data and information. For accountants, tasks such as data entry, financial statement preparation, reporting and invoicing will increasingly be automated thanks to the power of AI tools.
As part of FreeAgent’s wider research, we spoke with John Toon, Tech Strategy Lead at accountancy firm Beever and Struthers, to discuss his thoughts on where AI in accountancy is heading. He highlighted a potential future scenario that would save accountants countless hours: “You could have a client that comes to your website and wants to file their tax return. AI can gather their P60s and their dividends, file them in your document management system, extract the information and post that into your tax software. It can calculate the tax your client might potentially need to pay and share the links they need to pay it.”
Many time-saving capabilities are already built into the tools you use in your practice day to day, such as data categorisation (e.g. expenses and bank transactions). For example, FreeAgent has been using AI to help accountants and their clients save time and increase the accuracy of bank transaction explanations since June 2020. Using tens of millions of bank transactions, our data scientists adapted a machine learning model to accurately reproduce the way humans typically explain bank transactions in clients’ accounts. This way, AI-powered technology reduces errors and the time required to categorise client data.
Accountants should view the introduction of AI as an evolution of the tools already in place rather than as a replacement for the experts currently performing the task.
Reducing errors
Another significant opportunity for AI to bring value to accounting is within processes and workflows, especially when it comes to risk management and, potentially, even fraud detection as the technology continues to develop.
According to recent poll data gathered by FreeAgent (August 2023), accountants are intrigued by the potential impact AI and digital tools can have on reducing errors and future risks. When asked which tasks they would welcome AI’s help with, nearly two-thirds (65.3%) of participants identified preventing clients from entering incorrect information (e.g. using the wrong category for expenses) to be a priority.
In fact, identifying erroneous data and detecting anomalies earlier can prevent greater issues down the line. While some of these tools are still in their early stages, the rapid development of AI’s ability to help reduce errors is set to bring positive change.
Improving service
Increasingly, AI will empower accountants to focus on relationship-building, advising and offering guidance to clients. Skills such as empathy, collaboration and connection will carry higher value and greater focus since AI can support accountants on most data-driven tasks. As FreeAgent’s Director of Product Craig Clarke mentioned in a recent webinar (August 2023): “The promise of AI is that it will lift the burden of mundane, repetitive activity away from us to leave more capacity, context and care to tackle the challenges only the human condition can master.”
As a result, the proactive use of AI in your practice will free up time to establish stronger client relationships and more client-facing activities, such as consultancy services, or expanding your role as a strategic business adviser. For this specific focus, AI can help identify clients within your database that could benefit, for example, from a specific product or service, making your advice and guidance more personalised. Unsurprisingly given its nature, AI can help filter through variables and specifics to extract deeper insights into your clientele.
However, as we’ll see, while AI can provide support on many of these activities, AI-powered tools should be used with care: the individual should always be held accountable for making responsible choices with AI outputs, including security and privacy.
The potential of automation
FreeAgent’s Accountant Monitor survey shines a light on recent trends and issues in the industry. When asked about the opportunities automation could bring to their daily work, nearly 60% of respondents agreed on the advantage of using automation and AI-powered tools to spend more time offering tailored services to their clients.
Practical options for early adopters
AI has and will continue to develop incredibly quickly. It might not be something we are wholly aware of, just as electric cars, mobile touch screens and contactless payments gradually became part of everyday life. The software and technology practitioners use every day will increasingly be powered with AI functionality, meaning AI adoption in practice will happen almost subliminally.
That being said, there are many advanced tools that you can use today to elevate your practice processes in other areas. By harnessing these technologies, you’ll not only keep pace with industry advances but also position your practice to adopt future innovations.
Simplify admin tasks
One of the most actionable applications of AI is simplifying administrative tasks. Since its debut in February 2023, Google’s chatbot Bard has undergone several enhancements. Its latest release allows Bard to integrate with your Gmail and Google Drive and answer questions based on the information it finds. This means you can ask Bard to do things like summarise the contents of an email, generate meeting ideas or highlight the most important points of a document you have stored in your Drive. It can even use the information to create a chart or a bulleted summary. This simple yet highly effective development could help streamline workflows and create a more efficient workspace.
Boost your marketing
Marketing also offers an ideal entry point for AI. Platforms like ChatGPT and Bard have emerged as accessible and handy tools to assist content creation. They can help individuals or small teams to generate new ideas, optimise content for organic search and even refine messaging.
When it comes to video content, tools like Munch and OpusClip can be handy for identifying key moments in webinars and videos, making them ready for multi-channel sharing. From captioning and resizing and integrating with social media, these tools can simplify the process.
AI can also be harnessed for email marketing. Platforms such as HubSpot and Mailchimp analyse customer behaviour, segment audiences and fine-tune content to deliver campaigns for better engagement. And if you’re struggling with catchy subject lines to increase your email open rate, AI-driven tools like Encharge could be of assistance. It works in a similar way to ChatGPT, where your prompts guide the output.
Where AI still falls short
The human element
As impressive as these tools might be, they should be viewed as complementary to human creativity and expertise, rather than a replacement. We’re still a long way from developing an AI system that is able to manage and advise clients about their financial data autonomously. A tool like ChatGPT is a machine learning system designed to provide plausible responses to questions; it is not designed or expected to provide the same level of professional nuance and judgement as a trained accountant.
For example, accountants have an in-depth understanding of changing legislation and a genuine human connection that AI cannot replicate. Emotionally intelligent responses to client questions require soft skills, so while AI can be quick at analysing information and providing suggestions based on that data, human expertise is and will always be required to check and validate those suggestions.
As part of our wider research, we spoke with Russell Frayne, Head of UK Digital Accounting Solutions at accountancy firm Azets, who agreed that AI currently can’t offer the personalised human element that clients want. He said: “Clients come to us because there is a level of trust in our expertise. They want someone to look at their unique scenario and provide them with a personalised service. They don’t want something generic that’s been churned out from the internet.”
Accuracy matters
While AI boasts a broad knowledge base, it has its limits. For instance, at the point of publication, only those with Plus and Enterprise ChatGPT subscriptions can access a version of the tool that trawls the web for current information. If nonsubscribers ask ChatGPT about events after September 2021, it will draw a blank: “As of my last update, my knowledge was cut off in September 2021. I don’t have detailed information on events, developments, or changes after that date.”
While the company says it will roll out this feature to all users soon, it underscores the importance of human oversight - any AI-generated content should undergo review to check for its relevance and accuracy. It’s important to remember that ChatGPT and Bard don’t really‘know’ anything, but they are very good at figuring out which word follows another, which starts to look like real thought and creativity when it gets to an advanced enough stage.
Looking ahead
Addressing concerns
Recent and future AI advancements are set to have a positive impact. However, there is still much to be done before AI is fully embraced and implemented. Many concerns about AI – in particular around the security and accuracy of information – could prevent the adoption of AI tools and slow efficiency gains for practices. Largely, these concerns come down to what may happen if the technology is allowed to operate freely and independently.
While these concerns are valid, AI needs (and will always need) human intervention and expertise to ensure it works effectively. While everyday tasks can be drafted by AI, they must still be reviewed by staff and made customer-facing. To take advantage of the benefits that new technology can offer – particularly around streamlining admin – practitioners need to educate their staff about the crucial role they play when integrating with AI. Practices also need to have clear policies about how AI can be used safely and securely, such as guidelines about what data AI tools can access.
Similarly, concerns about job security can be addressed with better education. While AI is likely to reduce work in some areas of accountancy, it will potentially create new roles elsewhere that, with training, can be taken up by accountants who have previously worked in calculative roles. By showing staff that their work is still imperative, firms can promote the use of AI and redirect resources more easily.
“Clients come to us because there is a level of trust in our expertise. They want someone to look at their unique scenario and provide them with a personalised service. They don’t want something generic that’s been churned out from the internet.” - Russell Frayne, Azets
Responsible AI
For all its benefits, AI still needs to be used ethically and responsibly. Many of the world’s largest software companies and developers have agreed on the need for ‘Responsible AI’: a human-centred and ethical approach to AI to protect its users’ safety and privacy. In March 2023, the UK’s Department for Science, Innovation & Technology and the Office for Artificial Intelligence shared a white paper titled ‘A pro-innovation approach to AI regulation’. This policy paper included proposed regulations for the continued fair and responsible progress of AI technology in the country.
Final thoughts
Although AI’s integration into accounting is in its early stages, its momentum suggests it’s more than just a phase - it’s the future. For many accounting practices, the journey towards AI adoption is just the beginning, prompting them to evaluate and consider the initial steps towards integration.
For practices concerned about how they’ll pivot to an AI-driven landscape, it’s important to remember that the accounting profession has been evolving in the face of technological development for several decades. LLMs and GenAI are the next wave of that evolution, and while we may well be crossing a new and exciting threshold, the concept of using tools and systems to automate tasks is already widespread.
As AI advances, staying in the loop and incorporating its capabilities into daily tasks could equip your practice for the future. And while AI promises efficiency, it’s vital to remember that human expertise is irreplaceable. Striking this balance ensures an accounting practice that’s both cutting-edge and human-centred.
FreeAgent’s award-winning accounting software is powered by the latest technology, including AI, to make finances effortless for accountants and their small business clients. Find out more about FreeAgent for accountants.