Is AI really revolutionising the accounting industry?

There hasn't been a day this year without someone talking about AI. Chat-GPT made some waves in the accounting world when the info that it can now solve a sample ACA paper and fill in data for a US-based tax return hit the news. It may seem like AI is making massive strides into the accounting world (which is not entirely true) and we’re impending an AI-driven revolution. Is it really the case? What does it mean for your career and the future of your practice?  In this article we’ll look at the hype surrounding AI and help you manage expectations around it. 

Pavel Danilyuk // Pexels

All on board the hype train?

Using AI and ML algorithms in practice may seem scary when you think of it but realistically, you don’t need to be an expert. Don’t get wrapped up in the hype just yet. Dabble with it (with caution), automate stuff like email replies, data entry (not sensitive client info). Keep an eye on other industries and how they leverage ML and AI to get an idea of what’s possible. And read on because in the next section we will tell you a bit more about why what you think is AI and doing great could indeed be much better. 

AI is already present in your day-to-day

Chat-GPT aside, you may not even think about it, but you’re likely using some, albeit limited, form of AI every day. OCR tools have been using machine learning models for years now to speed up invoice processing. Human eye is fallible and attention span is limited, especially when looking at the same type of document for hours. While not always flawless (we’re looking at you Dext), OCR processing should still be more accurate than when done by a human and remove the need for triple-checking the same entry. 

But manual data entry is not the only space where artificial intelligence can supercharge accounting flows. Here are some other use cases for AI in your day-to-day: 

1. Real time data

AI and ML algorithms can cleverly aggregate and report on financial data flowing from different sources. Think revenue recognition from multiple income streams - both you and your client could have immediate access to real-time financial information about the business as a whole. This could otherwise be difficult to understand when looking at each stream of revenue separately. And cutting down time wasted on transferring data from one tool to another, the copy & paste and back and forth emails with your clients requesting reports, will bring you back the bandwidth for those meaningful conversations.

2. Smart processing and categorisation

Machine learning algorithms should make invoice processing and categorisation contactless. While you may be using rules and auto posting already, it still requires you to set it up. An ideal application would pull out more from the invoice without your involvement. We're talking category, project code, VAT and department or even bank details for future payments to cross-check with previous data. You should be able to fetch data from other sources and connections you have or automatically categorise if context is given in an email. Imagine for example that a receipt for a taxi is sent by Jane from Marketing. The AI tool could use the context given in the email (for example, her signature) to assign the spent to the marketing department. Smart invoice processing could take you from 10 minutes spent on an invoice to one click. 

3. Audit and compliance

AI works where the human eye fails. This could have a huge impact on streamlining tax return processes from preparation to the actual filing. While you can’t control how quickly your clients will get you the documents, you can control how well the data is processed. Letting an AI pass through the transaction and invoice history could bring out any issues before they escalate to a higher tax bill or late-filing penalty.

4. Fraud detection

In big firms, artificial intelligence could help with another aspect - fraud detection. Again, it can spot suspicious patterns before real damage is done. Something that may otherwise slip through the cracks.  

5. Practice management and admin

Summarising meeting notes, email chains, tagging relevant team members or even drafting emails and pulling the correct report to attach it to the message can all be automated. For the former use cases you can already use ChatGPT or other AI-powered editor tool, however you're still relying on the copy & paste to get it into your planning or client management flow. The recent announcement from Karbon AI is the step in the right direction for increasing productivity on the practice and client management side. Nontheless, we'd also like to see automatic emails with, for example AI-compiled list of missing invoices and that's something that's definitely on our radar at Aurelia.

6. Forecasting 

Last but not least, AI could learn about your clients’ spending patterns and revenue. This will give you high-level insights into how their finances will look in the short and medium term. It can't offer advice (that's your job!), but it can certainly provide some actionable insights. 

How humans and robots collaborate

Overall, AI and ML promise accountants greater productivity, accuracy and efficiency. Manual and repetitive tasks that require little human judgement can be automated and taken on by an algorithm that doesn’t get bored and doesn’t have limited attention span so can keep the speed and precision. In the long run this will have a positive impact on the bottom line - improved reputation and more clients equals more revenue. And who would say no to this? The key though, is to play your cards right. 

In order to successfully embed AI into your workflows is recognising the benefits and limitations of both human and artificial intelligence, and figuring out how these two can work together. 

AI is suitable for repeated tasks on big sets of data but won’t be able to recognise the intricacies of your practice and your clients' business. Unlike you, it won’t be able to engage in complex problem-solving. Humans not only have knowledge of tax regulations, practices and intuition that come from experience, but also empathy. AI still needs your supervision and input. Only when filtered through the lens of your clients' fears, motivations and goals, the aggregated, accurate, instantaneous data provided by AI will make sense. 

One more thing to consider is tech upskilling. Soon enough clients, especially millennials and Gen Z-ers, will be looking for more tech-minded accountants who really leverage AI in their processes, as they associate it with speed, accuracy and efficiency. Firms that pass on technology won’t be able to compete. Another thing worth considering is specialising. Knowing the ins and outs of a specific industry, like e-commerce, can be invaluable when high level accounting processes can be completed by a computer.

Final word

If you want to include power up your workflows with AI you need to recognise the benefits and limitations of both human and artificial intelligence, and figure out how these two can work together. 

AI is suitable for repeated tasks on big sets of data but won’t be able to recognise the intricacies of your practice and your clients' business. Unlike you, it won’t be able to engage in complex problem-solving. Humans not only have knowledge of tax regulations, practices and intuition that comes from experience, but also empathy. In it's current shape it's still not flawless needs your supervision and input. So only when filtered through the lens of your clients' fears, motivations and goals and the context of the business the aggregated, accurate, instantaneous data provided by AI will make sense. 

PS. At Aurelia we're leveraging AI to do smooth and smart invoice processing. We pull the data that's relevant with high accuracy and give you a confidence score to see if you still need to take a second look. In most cases, you don't, and the data extraction gets better over time. Intrigued? Drop us an email at [email protected], visit our website or come over to stand 590 at Accountex next week!

Visit our websiteor Chat to Jasper