How will AI soon help accountants work more efficiently? We discuss what’s on the near horizon and how your practice can take advantage of AI.
As we see artificial intelligence (AI) start to revolutionise the world around us, it’s natural to ask: how will AI impact accountants?
Ask 100 experts what AI is and you’ll get 100 different answers. What it’s NOT is a robot standing alongside the photocopier in each office. What AI is being used for is image recognition, object identification, detection, and classification, and automated geophysical feature detection – tasks that once required a human.
For accountants, AI will very soon help automate many routine and repetitive activities, but it will also:
- Empower quick decisioning
- Create smart insights
- Allow us to examine huge quantities of data with ease.
Here’s five examples of how AI will likely prove useful in accountancy over the coming years.
Free research report: The Practice of Now 2020
We surveyed 3,000 accountants from Australia and worldwide to reveal how the accounting landscape is changing. Discover how your fellow accountants are preparing for the next decade and learn what you can do now to keep your practice successful.
1. Predictive and forecasting solutions
Helping clients forecast their business finances is an extremely valuable service offered by accountants. With AI integrated into the solution, accountants will be able to provide comprehensive and accurate insight for customers without the usual ‘manual heavy lifting’ and number crunching. On a day to day basis, being able to quickly and easily access up-to-date and accurate reports and forecasts can help an accountant form a closer and more useful client relationships.
This revolution will be empowered by one of the cornerstones of AI today: machine learning, which is the ability of software to essentially program itself based on the data it encounters. The software can learn from what you do with data and make its own suggestions for humans, if not act entirely autonomously.
Machine learning is everywhere. It makes smart phones possible by enhancing predictive text, making speech recognition possible, creating route suggestions when navigating, or suggesting a place you might want to visit at your destination.
In organisations worldwide, 77% of businesses already say they’re completely or very reliant on machine learning technologies (source: Cisco Systems 2018). Other research says the top use for machine learning currently is more extensive data analysis and insights (45% of respondents; source: Technology Review and Google).
What machine learning needs—and what makes it so useful—is access to data. Lots of data. That’s why machine learning is coming to the fore now, because technology like the cloud means all the data can be collated and is accessible, rather than hived off within discrete systems. Cloud computing also means we’re able to generate more useful data.
2. Smart assistants
There are probably accountants reading this blog who, during crunch time when seemingly every client is sending through their accounts, thought to simply turn off the phone or email so they can get some work done. Well, smart assistants might just take you closer to that goal. They can form the first line of customer contact and provide customers with the information they need, such as details about their current tax liability.
Lots of us know about smart assistants because we interact with the likes of Siri or Amazon’s Alexa. Sage has its own smart assistant called Pegg that can be used for client accounting. People can ask how much money is in their payments account, and Pegg will tell them. They don’t need to understand accounting terminology, or even what a ledger is.
Smart assistants come in two forms: natural language bots, and scripted bots. Scripted bots have been around for a long time—they’re easier to build and mostly used for mobile engagement strategies, so you might encounter them on a website. They simply recognise key phrases and aim to provide a ready-made response. Sometimes these are called chatbots.
Smart assistants are an order of magnitude more sophisticated, and often involve speech recognition and accurate human voice synthesis, so they can respond to natural language queries. Smart assistants also learn the more you use them.
Both smart assistants and scripted bots have their uses and one shouldn’t be considered better than the other from a business perspective.
3. Automatic tagging and allocation of transactions
The next two areas where AI will help accountants are also enabled by machine learning.
Machine learning will save time for accountants by correctly tagging transactions and assigning to the right ledger account. Put simply, it learns from previous tagging decisions typically made according to rules the accountant is aware of. Some of these rules are intuitive but some can be surprisingly complex—at least from a computer’s point of view. Over the coming years the ability of technology to discover these rules and predictively plan will help remove a significant component of the accountant’s daily drudgery and workload.
4. Anomaly detection
Computers love data, of course, and when machine learning is applied to massive amounts of data—such as the yearly ledgers of a large company—there are clear benefits. Accountants will be able to discover anomalies much quicker, with significantly reduced effort compared to previous methods.
For example, if an audit is required it will be possible to audit all the data rather than merely a sample, yet without the huge resources typically required for what’s traditionally considered a “full” audit.
5. OCR solutions
Optical character recognition (OCR) is not new but AI enhances its accuracy significantly—and opens it to new uses. While it’s previously been possible to extract information automatically from documents, this required a human to point out to the OCR software where the data was located—something that also meant the document layout couldn’t alter without further instruction.
Computers have always known what numbers are, of course. That’s what defines a computer. A printed receipt for a purchase is full of numbers, but they’re certainly not all equal. Some are of particular importance to you as an accountant: the date, the total amount, and perhaps the credit card number. A human can instantly identify these numbers without even thinking, but until now they were indistinguishable for a computer. The digits 1-5-1-2 might be the last four digits of the card number, a date, or the amount of an item on the receipt.
With the application of AI to OCR, the OCR software can recognise printed documents like receipts or invoices, on which it also recognises and extracts salient data. Data can then be allocated and/or processed by the software rather than a human, even if it hasn’t seen a similar document before or the scanned document isn’t high quality. This reduces the human effort and time needed to assign information.
So what’s next?
How will we get all these wonderful AI-enabled benefits? The terrific news is you probably won’t have to do very much! In the old days we might’ve expected to buy add-on software to gain revolutionary new functionality. Now, the kinds of tools discussed above will more than likely simply start to appear in the software you already use over the coming years.
Some of it, such as bank account reconciliation, might already feature in your firm’s accounting and client management software—and you might not even be aware. All you might’ve noticed is things got a little bit easier when the software appeared to get that little bit smarter. This is the shape the revolution will take—many small increments, rather than an overnight change.
There’s always a “but” in stories like this, and here it boils down to this: AI is part of a wider revolution in technology that’s enabled by the cloud. Cloud computing is the only way to collate and make freely available the massive amounts of data machine learning needs, as just one example.
We mentioned above how machine learning might become the best auditor in the world and spot errors humans struggle to see. However it can only do this if all the data is accessible. If the data it requires is spread over 100 spreadsheets or, even worse, printed documents, then it simply isn’t possible.
In other words, AI builds on technologies like cloud technology. The message is simple: if you want AI then ensure you’re taking advantage of the latest technologies. This doesn’t mean sitting on the bleeding edge of technology, but simply being aware of what’s common-sense good practice as far as technology goes, and ensuring it’s adopted in your practice.