We discuss the impact of artificial intelligence (AI) on the tax and accounting profession, from key trends to job security to how AI will change your day-to-day workflow and more.
The impact of artificial intelligence on our profession
In the first episode of Thomson Reuters Backstage, Amanda Trent, Segment Marketing Lead in the Thomson Reuters Digital Demand Center, explores the impact of artificial intelligence (AI) and all it has to offer tax professionals.
What are some of the key trends for AI in tax?
AI is a broad form of technology, what exactly will be the impact of artificial intelligence on you and your firm? According to Jordan Kleinsmith, Product Innovation Lead for tax and accounting professionals at Thomson Reuters, the major AI trends for tax and accounting fall into two areas: aiding staff to do their existing work more efficiently and allowing them to explore new services they can provide to clients. AI-enabled programs allow staff to process more returns in less time as it can help to answer questions on complex topics faster. When staff have this extra time to spend with clients alongside artificial intelligence, they can uncover new ways to serve clients thus helping to maximize both the firm’s top and bottom line.
Am I going to lose my job?
If you’ve been following the Thomson Reuters tax blog, you’ll know that this is not the first time this question has arisen. It was also covered in a previous post, “Are employees being replaced by artificial intelligence?” It keeps coming up because it’s usually top of mind for tax professionals. However, you can rest assured that your job is not likely at risk. In short, “we’re not really attempting to create a bunch of accounting androids, we’re really looking to create accounting cyborgs,” says Jordan. The collaboration between humans and artificial intelligence is what makes this technology so powerful, not just AI alone. According to Khalid Al-Kofahi, who leads The Center for AI and Cognitive Computing at Thomson Reuters, the goal is to minimize the number of menial tasks that accountants have to do and enable them to spend more time doing the part of their work they enjoy most, working with clients.
Just think about how social media has revolutionized the workforce today — people were much more likely to see their jobs change with this technological advancement rather than losing them. If anything, more jobs were created in response to this new technology. It is anticipated that the impact of artificial intelligence will be similar.
You may be wondering why this pressure to switch to AI-enabled programs is happening now. In reality, companies have been utilizing this technology for quite some time. While the accounting profession may be averse to change, the pain points facing industry professionals are not going away. In order to remain competitive, it is imperative that firms adapt with technological changes. By nature, one impact of artificial intelligence is that it increases its capabilities and knowledge with time so the longer you wait, the farther behind you will be.
On top of this, client expectations have increased quite dramatically. Today, 41% of clients expect an answer in less than six hours, according to Pushkar Bhoopalam, the Head of Research Suite of Products. To meet this expectation, firms need to have a better, more reliable way of finding the answers to these questions. To make matters worse, Pushkar mentions in his conversation with Amanda Trent that “it’s a reputational risk when you get something wrong these days.” Firms can’t afford to make mistakes, especially with the speed in which information travels through channels such as social media.
The timing is also right because Checkpoint first launched in 1997. And this, as Pushkar points out, makes it 22 years old — old enough to be taken out for a drink if it was a person. And when you ask an adult a good question, you expect a good, coherent answer. Checkpoint should be the same way as well. This gap provided the opportunity to revitalize the product to better serve the customers of Thomson Reuters, so then they can better serve their own clients.
What will AI in my everyday work look like?
A very common impact of artificial intelligence in the tax and accounting industry is in research. The Tax Cuts and Jobs Act (TCJA) brought about almost 3,000 of new pages to the federal tax code alone, not counting that of the individual states. Expecting a CPA to understand every aspect of all of those changes, especially those with clients across multiple states, would be silly. Instead, we can enable staff to find the answers they are looking for in a much easier way. Rather than typing in keywords and sifting through information, staff can simply type in their question, in plain English, like one would with the Google search engine. These sorts of features help to make a user-friendly system, which is what AI is all about. These things help the right things to be found in much less time, and your firm doesn’t have to worry about the information’s credibility as it’s coming from only the most trusted sources (i.e., the IRS, AICPA, EY, Deloitte, etc.).
For AI to be more widely used by tax and accounting firms, it needs to become as accessible as possible. In the beginning, only the large firms that had the in-house abilities worked with artificial intelligence. But now, as more software available for purchase such as Thomson Reuters Checkpoint Edge, smaller firms can put their hat back in the ring and the overall industry will benefit from the increase in quality.
How does AI work? What makes it different than other existing technologies?
One of the aspects that differentiates artificial intelligence is the user’s ability to engage in more natural dialogue with the technology. Instead of engaging in a constant back and forth search process, AI uses predictive search to help get to the question you’re asking and put that best sources on top of the search, all so that you can get to the information you’re looking for in less time. The technology can also suggest concept markers of different topics the searcher may also be interested in learning about which can eliminate another separate search in a lot of situations.
Other frequently asked questions
Since this webcast originally aired lived, viewers were given the opportunity to ask the aforementioned subject matter experts any questions they had. The following includes Khalid Al-Kofahi’s answers to some of the most frequently asked questions that were received.
Will accountants need to learn how to code?
Learning to code while fun is certainly not required. One of the key objectives of AI and machine learning is to shift the burden from the user to the machine and to deliver more powerful capabilities in simpler and more intuitive experiences. This is true not just for AI but in other industries. As technologies mature, toolboxes become appliances that just work. We have a long way to go before this becomes a reality in Tax software (it just works, no intervention needed), but that is where we are going.
Will the AI created be similar to an “Alexa” or “Siri” where one can dialog and ask questions, etc. And if this type of AI is implemented, would this be for transaction entry (i.e., speaking an invoice into a system and it being auto-entered)?
Alexa and Siri (and other digital assistants) are popular for a number of reasons, but primarily because of their convenience; they are simple to use do a pretty good job on the tasks they were programmed to do. In our view, these are key requirements for successful applications for tax professionals as well. Simplicity almost always includes interacting in natural language. Whether it is asking questions or expressing complex information need in a natural language query. State of the art in AI and natural language processing is not capable of supporting open-ended dialog between human and machines, but within the confinement of carefully scoped vertical experiences some forms of dialog (e.g., allowing a customer to ask follow up questions, to refine or elaborate on previous queries) is possible and is an active area of exploration for us. Also, I am not sure if reading an invoice results in a better experience than just scanning it into a smart application that can parse the scan or picture and fill target fields automatically.