How AI Will Change the Face of Accounting

7 min. read

Accounting: few words inspire such dread in the world of business! Keeping books updated is a dreary and time-consuming activity. Fortunately, this task has become less burdensome with the expansion of software like Quickbooks or Pennylane, which centralize data while introducing the industry to newer, user-friendlier tools. Though most accounting software already uses some AI tools, generative AI is set to bring impactful and game-changing functionalities. In this post, we will review how artificial intelligence will change accounting practices and facilitate the task of the accountant.

How Can Accounting Profit From AI?

A rapid look at past practices shows us the extent to which technology has changed the work of an accountant. Initial software transferred accounting work from paper to computers, enabling a first wave of automation.

Software then progressed to become a decentralized service hosted off-site. All the data silos were suddenly brought down, enabling companies and their accounting firms to access and review their accounts in real-time.

This evolution to SaaS has also brought more flexibility to the development of software. To maintain their cutting-edge position, the new wave of SaaS companies has to innovate. That’s why they have been quick to adopt forms of AI and amongst the first to profit from the new AI revolution.

The first AI wave: tools to ease bookkeeping

Optical Character Recognition, or OCR, was the first AI tool introduced into accounting.

The most time-consuming accounting task is journal entry.

OCR can identify fields like the name of the supplier, the total amount, or the VAT, and pre-enter them, only leaving the accountant the task of checking if this is correct.

According to Thomson Reuters, other ways accounting firms use AI-powered tools include client communication with chatbots, data analysis, and tax research.

Nevertheless, AI is still restricted to the automation of specific tasks, more specifically those relating to the first layer of accounting ie. bookkeeping. The new wave of generative AI is the real game-changer.

The generative AI wave: a copilot able to search and execute tasks

Ramp exemplifies the next step for this kind of AI tool. Not only does it help streamline bookkeeping and automate review processes, but it also introduces new ways to interact with the accounting data in natural language.

This is the main development brought by generative AI: it acts as a helpful assistant capable of answering every inquiry.

The copilot can browse the accounting data, sum up invoices, classify them according to their status (to look for unpaid ones for example), or help manage expenses.

Here’s how accountants can offload work onto an AI copilot:

  • Help answer requests from clients: every day, accountants receive numerous requests and queries from clients. Most require browsing through entry data. Instead of spending 10 minutes finding a thorough answer, accountants can simply ask the copilot to retrieve the right data.
  • Automate tasks: as we saw, AI can already help interpret invoices or receipts. The next step involves being able to completely automate data entry and produce reports when asked.
  • Provide clients with simplified access to their accounting books: some accounting software involves steep learning curves. Getting the right info can require some time. With an AI copilot, accounting firms can give their clients access to their accounting book, without without them having to delve into the software.

What are the use cases AI can run in accounting software?

We asked numerous accountants and accounting software providers about the key use cases they identified as critical and which can be powered by AI. Here are the ones most frequently mentioned:

Look for unpaid customer invoices

Tracking unpaid invoices is vitally important for companies of all sizes. A study made by Xero shows that small business owners spend around 10% of their time tracking unpaid invoices with a 14-day average between the date of the emitted invoice and the actual payment.

An AI copilot connected to your accounting data can help you retrieve those invoices and better figure out where money is missing, and who owes it.

Get insights on supplier invoices

One way to keep expenses under control is to know exactly where the money is going and for what purpose. That’s why some find tagging invoices crucial:  it helps keep budgets under control, and if need be, pinpoints possible savings.

Artificial intelligence can help supply those insights: for example, you can ask for the total amount spent for a given category, or supplier.

This helps track expenses for suppliers whose invoices are based on usage, like AWS.

Provide personalized financial ratios

For most business owners, the most important things to know are summed up by key financial ratios like the current Net Income, or Cash ratio.

As such, those ratios are not hard to calculate, but getting the freshest data is more tricky and necessitates a call or email to the accountant.

A Copilot can provide real-time ratios based on the latest transactions.

Managing taxes

Juggling between tax regimes can be hard to follow, especially VAT. And in the US, it can be even harder as each state has its own set of taxes.

Did you know that in France, you can’t reclaim VAT on business presents that are worth more than €69? VAT regulations are full of such particularities, making it hard to keep track of information.

An AI copilot can provide you with valuable information: when to pay, the total amount to be paid and what can be deducted.

Will AI steal my job? And Other Legitimate Questions

Still, some friction remains. The main friction is inherent to the sheer novelty of this technology. Most accounting firms remain wary of using AI because the technology is too novel, perceived as not sufficiently secure, sometimes misleading, or positively wrong.

The way generative AI is impacting the nature of accounting does create another type of friction: will ChatGPT steal my job?

Bundled together, those assumptions create understandable concerns, which hamper the wider adoption of powerful AI tools.

Let’s review those concerns.

AI isn’t new: most accounting software already uses it!

Though generative AI is a new technology, AI as a whole isn’t: OCR, for example, is an accounting software staple and has been for the last decade or so.

There are now multiple layers of AI in most software, such as machine learning. They learn from user’s usage to improve results, helping to automate tasks or detect anomalies. Pennylane, for example, uses machine learning to optimize its labeling system.

And some of the market’s leading software providers, like QuickBooks, are developing copilots on their side too.

AI is secured because it needs to be

A big concern about generative AI is security. Lots of businesses are wary about their internal data spilling out into the open, and they fear that using an AI tool like OpenAI will just do that.

But OpenAI has pledged not to use data shared with its API, and its partnership with Microsoft has led to Enterprise plans with Azure, Microsoft’s cloud solution which securely hosts the data of tens of thousands of companies.

Other companies providing services similar to OpenAI adhere to the same level of security. Anthropic is now backed by Amazon — which owns AWS, the biggest cloud service. Mistral, a French start-up has pledged to use only open-sourced data, etc.

The bottom line is that generative AI is safe because if the companies providing AI want to sell their services to businesses, they have to comply with top-notch security requirements.

A 95%+ level of accuracy

A generative AI tool like ChatGPT alone won’t provide the best result: only tools allying large language models like ChatGPT, Agents calibrated for accounting use cases, and data can deliver the best results.

For example, at Blobr, we leverage the APIs of your accounting software to execute tasks.

We then create agents aimed at delivering specific use cases through different channels.

The final product is a super-focused AI copilot for accounting that guides the generative AI to perform precise tasks, leaving next to no room for error.

No, AI doesn’t mean the end of the accountant!

Losing jobs to AI is a widely — and legitimate — worry. Why should companies still use humans when machines can perform the same tasks? OpenAI itself sees accounting as one of the most jeopardized industries.

Yes, AI will redefine accounting practices and an accountant’s daily work. Most tasks will be automated, and already are — OCR is an example of AI streamlining book entry. But those tasks are the least “value-loaded”, and the most time-consuming.

Accountants are aware that their industry is impacted by a structural workforce shortage: fewer people are studying accounting, while lots of accountants are retiring. The workload keeps growing, and the workforce keeps shrinking. According to a study by FloQast, this has led to an astonishingly high percentage of accountants — 99% — experiencing some level of burnout.

Aside from alleviating the workload, AI can redefine accountants’ relations with their customers: a copilot can handle simple requests, freeing up valuable time for accountants to better help their customers.

Forget the cliché of the accountant with their nose stuck in Excel figures. Human relations are an important aspect of an accountant’s daily work. Their knowledge is invaluable in an increasingly complex and regulated environment. Business owners are increasingly seeking their advice.

AI can free up time for accountants to give advice and respond to their clients’ inquiries.

The advent of AI in accounting is not a doom-and-gloom scenario where machines replace humans. By automating routine tasks and providing intelligent assistance, AI enables accountants to focus on the more complex, strategic — and arguably satisfying — aspects of their role. The future of accounting is not just about numbers — it's about humans and AI working together to achieve greater efficiency, accuracy, and strategic insight.

Stop wasting your Ad budget

Book your 30 minutes call to setup your goals and receive your first report.

Keep reading

October 12, 2023

AI Chat, AI Assistant, AI Copilot, AI Sidekick… Your Guide to New AI Definitions

AI chat, AI assistant, AI copilot, and AI sidekick... They re all AI-powered chats, but they don't bring the same kind of services. Learn the difference between them.
October 9, 2023

How AI Will Change the CRM Industry

AI tool can be a real game-changer for CRMs, enabling users to search and perform complex tasks, all in a simple conversational interface. In this piece, we will define and review AI use cases for CRMs and compare the main existing AI copilots.
October 18, 2023

The Guide to Large Language Model APIs

In this article, we explain how to use the LLM parameters and how they work, while benchmarking the available LLM APIs.