In this podcast we are joined by Cory Ng, associate professor of instruction at Temple University in Philadelphia, and John Alarcon, founder and principal of BEARN, a management consulting firm in Philadelphia, to discuss their book, Artificial Intelligence in Accounting: Practical Applications. The authors, who are also Pennsylvania CPA Journal Editorial Board members, look at artificial intelligence’s current usefulness within the profession, the importance of robotic process automation, and the challenges and considerations CPAs will have to face when adopting this technology into their processes.
By: Bill Hayes, Pennsylvania CPA Journal Managing Editor
In their recently released book, Artificial Intelligence in Accounting: Practical Applications, Pennsylvania CPA Journal Editorial Board members Cory Ng and John Alarcon seek to help CPAs to better understand artificial intelligence and its many uses in business and accounting. Ng, an associate professor of instruction at Temple University in Philadelphia, and Alarcon, founder and principal of BEARN, a management consulting firm also in the city of Philadelphia, join us today to discuss the subject matter of their book and its value to the CPA world.
Where did the idea of writing a book on artificial intelligence in accounting come from and how did you decide to write it together?
[Ng] The publisher, Routledge, approached me about writing a book focused on emerging technologies primarily because of my teaching and research profile. And I immediately thought about artificial intelligence because we're all connected to artificial intelligence in our everyday lives. If you think about using virtual assistants on our iPhone, or a device like Alexa, we're using artificial intelligence every day, and the implications for the profession are really profound.
That was the initial idea: writing a book about an emerging technology. The most applicable one that I felt was relevant for accounting was artificial intelligence.
[Alarcon] We also did some articles together, and one of them was on artificial intelligence, which was a good foundation to start that bigger project of putting a book together.
I was honored that Cory thought of me to be a coauthor for this book. We've known each other for a very long time. We're both CPAs with doctoral degrees and we have similar research interests. Cory works in academia. I work in business and industry. This was a great opportunity for collaboration between the world of academia and business for such a timely topic that everyone wants to know about.
There's a lot of information out there about AI, but it's all over the place and it's hard to get your head around it. Really, my primary motivation for participating in this book was as a way to bring all these parts together from the perspective of what AI means for accountants and financial professionals in general.
[Ng] Like John said, we've known each other for a really long time, more than 10 years. Both of us have served on the Editorial Board of the Pennsylvania CPA Journal and we've coauthored multiple articles on different technologies, artificial intelligence being one of them. So just my association with John through the Editorial Board, and also through his involvement at Temple University as an adjunct instructor, I just thought John would be an ideal person to work with and I'm so thankful that he agreed to coauthor this book with me.
What is it about the current state of the profession that made it important to write about this subject right now?
[Alarcon] The first thing that came to my mind when considering cowriting this book was to address the question out there on whether the accounting profession is going to survive this technology revolution, and, if it does, what it will look like in the world of hyper automation. Think of hyper automation as the application of advanced technologies like robotic process automation, artificial intelligence, and machine learning.
When we refer to AI, we define it broadly as a computer program or a system or a software application that can imitate or simulate human behavior; in some cases, even exceed human performance. We are talking about software programs and algorithms that can take on complex tasks or problem resolution.
These algorithms are used by software developers today. They've been around for a long time and they get better and better, but they are available in libraries of language, programming languages out there, or part of platforms, or even sometimes they are embedded directly in applications such as data mining, document management, or business application, which means you don't need to be a programmer or technician to get the benefit of AI. It gets delivered to you via some of these applications clearly.
Interestingly, at the same time that AI is becoming more accessible and taking over labor-intensive tasks of accountants, the profession is in transition. We are clearly seeing today a shift of the profession from backward-looking financial statements to more forward-looking financial performance management assessment and decision support. Really, the way to look at AI is AI is empowering accountants to achieve that vision.
[Ng] We are currently experiencing a digital transformation that's often referred to as the fourth industrial revolution. Artificial intelligence is one of the key technologies driving this industrial revolution.
Like John was saying, accountants need to understand artificial intelligence, the role that AI will play in business and accounting going forward, and how to leverage it in order to create value. Accountants have a track record of embracing technology to be more productive. If you think about prior technology innovations – the PC revolution, the age of the internet, cloud computing – accountants have consistently demonstrated their ability to embrace new technologies and incorporate them into their processes and services.
This is the next frontier for accountants. This is the next frontier in the digital revolution. If accountants understand how this transformative technology is reshaping the business landscape, the accounting profession is going to not only survive, but they are going to thrive by embracing AI.
The goal of the book is to provide accountants with a foundational understanding of artificial intelligence and its many business and accounting applications. What would you say are one or two of the most important business and accounting applications of AI as it pertains to CPAs?
[Alarcon] The two areas that I can think of immediately is accounting automation and auditing. Accounting automation because we are going to enter into a new era of hyper automation where mundane tasks are going to be replaced by software, data entry, reconciliations, inventory counts, and so on and so forth, but also auditing because AI will not only automate mundane tasks, but also analytical tasks with machine learning and deep learning.
[Ng] One of the most important applications for accounting for me was also in auditing, and perhaps I'm biased because I used to be an auditor, but that was where I saw the first huge impact. In Chapter Two, we discuss different ways that artificial intelligence can be applied. It can be applied in all phases of the audit, from planning to data gathering to completing the audit and issuing the report.
There are specific examples that we can look at. Neural networks have been used, for example, for performing analytical review procedures and risk assessments. AI algorithms can also be used to classify accounts receivable transactions, so whether a receivable is collectible or if it's bad debt. Then, in Chapter Five, we actually interviewed global professional services firms. There's one case study there that talks about AI-enabled autonomous drones being used for inventory management.
It would be pretty much impossible to fully explore some of these larger issues in the time that we have, but I figured I'd throw out a few and you could give me a brief explanation of their importance and their applicability to CPAs, if you could. How about natural language processing?
[Alarcon] That's a good one. Natural language processing, also referred to as NLP, is a subfield of AI that really focuses on the interaction of computers and people using human languages. An example of that that everybody thinks of is Siri, the Siri application on Apple's iPhone, or Amazon's Alexa, both use NLP to understand requests, to complete a given task, which, for example, could be recording an item on your shopping list or placing an order.
Now we are seeing early uses of NLP in accounting and tax as we speak, where software interprets large volumes of unstructured text, such as the tax code, for example, and can answer questions and make recommendations – the software – based on a particular situation.
So, if you will, accountants are now augmented by NLP engines to make sure they don't miss a deduction or a tax credit one of their clients might be eligible for. That's the kind of use we are seeing already today in a real-world application.
[Ng] NLP is a great illustration of application of artificial intelligence by the accounting profession. In Chapter Five specifically, we discussed the case study that explains how a global professional services firm is using NLP for risk analysis. As a result, they're saving a lot of time using this technology. And essentially, NLP can be used to review a large volume of contracts and other documents that are needed to evaluate the riskiness of loans, but they're able to do it in a matter of hours by using NLP, as opposed to taking several weeks if humans were doing the same task.
As you mentioned earlier, you've done lots of articles and features for the Pennsylvania CPA Journal. People can certainly check those out as well on the PICPA website. Robotic process automation, I believe, you guys covered for us. What's the importance of robotic process automation in the accounting industry and how do you discuss it in the book?
[Ng] We define robotic process automation as a software application or robot or bot that automates a business process by replicating the actions of humans' performance tasks within digital systems, such as manipulating or transferring data. If you think about all of the actions that you take using a computer, for example, to open programs and transfer data and send data to someone, all of those steps could potentially be automated using RPA.
RPA lends itself well to high-volume and repeatable tasks traditionally performed by humans. For example, RPA can be used to process bulk transactions such as processing vendor invoices for payment. In Chapter Three, we discuss a distinguishing feature of RPA as the ability to work across different platforms, so legacy ERP systems, mainframes, desktop applications, and others.
If you think about Excel, if you're familiar with macros in Excel, you can automate several different tasks within Excel, but you can't use that same macro outside of Excel. The nice thing about RPA is it's a technology platform that can be used across different platforms. Any platform that a human can use, potentially an RPA bot can automate those same tasks that a human would do working across different platforms.
So in Chapter Three, I referenced an example from Deloitte in which RPA is used to open up emails and attachments, and then connect to various systems, then extract data from documents, such as invoices, and then perform the calculations, transferring data, and much more, all of which can be done automatically by using RPA.
[Alarcon] The nice thing about the work that Cory has been highlighting is the fact that we provide a lot of definitions and illustrations so that you can understand how these different technologies interact together, what they mean, and we even give examples of products out there with examples of use of these products with actual companies, and all in one place, in one book.
What I would add to what Cory just said is there are some people out there, some experts that consider RPA not really part of AI; however, the technology enables machines to perform functions usually carried out by humans, and it is now coupled with AI capabilities, such as machine learning. We see these RPA systems that are AI-enabled specifically totally part of AI as we understand it.
How about what's known as text mining? What will be the impact there?
[Alarcon] Text mining is really booming right now in terms of technologies, progress, and innovation. Think of the information that accountants are processing: how much of that data is documents or text? There are a bunch of texts and sources that accountants have not been traditionally using just because of the volume of that information, which now, with technology, they can incorporate as part of their audit program.
It's super exciting. Text mining is one of the most exciting areas in AI. Again, like RPA, text mining in itself is not considered a form of AI. It is really a part of data mining, but the technology is now coupled with AI, which makes AI-powered text-mining applications part of AI. By text mining, we define it as a process, a process and technology that basically extracts relevant information and meaning from unstructured textual data from Word documents, PDF files, social media posts, emails, webpages, XML files, ATC filings, so on and so forth.
An example of use of text mining in the accounting profession today is contract analysis. There is contract analysis software that currently automates the analysis of legal documents and contracts to extract relevant information from them and compare them with guidelines and automatically verify compliance with standards or regulations.
An example of that is the automatic review of customer contracts and systematic review of customer contracts to ensure compliance with revenue recognition standards, for example. Timing of revenue recognition is a big issue for companies, and when you have a large organization with thousands of contracts, and traditionally accountants and auditors have been using samples, but now they can actually review the entirety of the population of customer contracts to detect any potential issues around compliance with revenue recognition.
Another example is vendor contracts. Vendor contracts – the use of text-mining software in vendor contracts – is really something that is going to take off dramatically. It's going to help procurement departments ensure vendors’ contracts comply with policies, and also things like legal compliance requirements, whether these contracts, for example, comply with security, IT security or data security. Do the vendors have proper language as it relates to data security or privacy, GDPR, and others?
Another example that we've seen in the press in these last few years is the use of these contract analysis solutions to review lease agreements to facilitate their review to ensure compliance with the latest lease accounting standards.
[Ng] Text mining really helps analyze patterns and trends and themes from unstructured text data. Whereas traditional data mining does this using structured data like in databases, text mining is used primarily for unstructured data. Like John was talking about, all those documents, contracts, and other legal documents, you can use text mining to understand these patterns and understand these themes.
It's a really useful tool. When you couple that with artificial intelligence, it can have a really pervasive effect on the accounting profession. The book discusses many applications of text mining in accounting, in auditing, and tax and advisory services.
One thing to realize is that, with text mining, now accountants and auditors will be able to process a lot more data a lot faster than ever before.
What would you say are some of the challenges and considerations CPAs and CPA firms may have to face as they look to move forward and adopt artificial intelligence into their processes?
[Ng] There are many, and we dedicated an entire chapter of this book to the challenges faced by the profession with AI. Chapter Six. One of them is algorithmic bias. If you think about it, humans are the ones that create the algorithms. If humans are biased, then that can lead to biased programming and biased algorithms. A biased AI system can literally lead to intentional or unintentional negative consequences or inaccurate results. That can be very concerning. There are a lot of ethical considerations for the accounting profession and we go into great detail with those.
[Alarcon] Algorithmic bias can be found, as Cory said, in the algorithm itself, but also in the training data that goes with it, and it can also reside in the input information provided to the system for processing.
So you have to really be vigilant in terms of risks for biases from these three aspects, which is basically input into the algorithm, the algorithm design itself, and the training data, which is basically the base that the machine learning system is using to learn from processing the information.
But the algorithmic bias is not the only challenge out there for accountants. We also talked about issues and challenges around security and privacy. Think about it: At the time where auditors and other accountants and other operators can access vast amount of data from emails from employees or customers or whatever, social media posts, you have to be careful that you don't violate any regulations.
There is clearly efforts right now to provide guidelines for technology providers and service providers to ensure proper security controls around these systems. Currently, there's still a lot of work to do and that's going to be the major challenge for the users of AI technologies.
In addition to security and privacy risks, you have the issues around change management. Look, these systems are more complex than ever before. The logic. The training data. Processes are going to be more and more data-driven with things that are generated automatically, and you have the issue of maintaining this automation when there are changes in standards or regulations, for example, which happen frequently in the world of accountants and tax professionals.
That's going to require major resources. Usually, companies going into AI projects underestimate the resource that will be needed to maintain their systems. In the book, we talk about it using examples in RPA implementations, but in other areas as well. This is going to create a lot of opportunities for accountants by the way, because they will be needed to assist IT professionals in the maintenance of these systems to make sure they are timely and accurate in their predictions.
What would you say is the main takeaway you would want CPA practitioners to get from this book? How will it help them in their careers and their development within the profession?
[Ng] I think it's pretty clear that AI will bring tremendous efficiency gains, time savings, and opportunities to increase the quality of accounting services. Accountants that have a basic understanding of how this disruptive technology works and its many practical applications that we try to highlight in the book should help accountants be able to hopefully identify opportunities where AI could be used within their own organizations, and then ultimately result in improved productivity, improved quality. There are potential applications for advising clients on how they can use artificial intelligence.
Just thinking about how it could be used to benefit internally organizations, and then those clients that you're serving is just really critical. Accountants can help future-proof their careers by understanding this technology and using this technology to solve business problems and create value.
[Alarcon] I think my takeaway from the book is really AI is going to be pervasive. The job of accountants is not going to go away. It's going to be impacted dramatically with higher quality and timeliness of services thanks to the technology. An example of that is because of their ability to process such large volumes of data quickly, you're going to see the processes of auditing change or evolve over time, become more timely and richer in insights for the organization.
We are also going to see new job opportunities and services from accounting professionals emerge. Examples of that are going to be around the design and the maintenance of accounting algorithms, as I was discussing earlier, but also auditing AI systems. You're going to need accountants with technology skills to be able to audit these systems to make sure they are free of bias and also they are accurate or compliant with whatever regulations.
Another example of an area where you're going to see a boom in demand for accountants is going to be the review of internal controls and processes and systems that reply on AI algorithms, which are going to be a lot of them because AI is going to be pervasive, touching every part of the organization from customer relationship management to production, planning, supply chain, and accounting back office, front office, the entire enterprise. There'll be a desperate need for people who are able to review and make recommendations for adequate internal controls around these processes moving forward.
The one thing I would just add is, as a result of all this, I think that that's why this book, what is important is we have a responsibility to not only embrace the new technologies and transform the profession, which we're all working on, but also prepare the new generation of accountants out there to meet the challenges of the profession moving forward.
[Ng] Our book represents months of research by John and me, assembling disparate and sometimes outdated information out there about artificial intelligence. Our hope is that, for all accounting and financial professionals and educators and students, this book will be a great resource for all of them as they think about the future of the profession, especially for those students that are, whether you're in an undergraduate program or a graduate program or getting ready to enter the profession, you'll have a leg up on everyone if you have an understanding of these technologies and its potential applications.
As part of my role in academia, I've already been incorporating data analytics into accounting courses, and I've been mentioning artificial intelligence, but my plan going forward is to further integrate artificial intelligence concepts and talk more about applications, and, in fact, incorporating this book into a graduate course that I'll be teaching at Temple University in the spring.
Excellent. Cory and John, thanks for being here with us today to discuss your book and artificial intelligence's importance in the accounting landscape. Certainly thanks for everything you do for the Pennsylvania CPA Journal Editorial Board. It's been mentioned here that you guys have done a bunch of work for that's been helpful to the accounting community, so I don't want to overlook that either. The publisher is Routledge, so feel free to check it out on their site. And when in doubt, feel free to simply Google the name of the book, which is Artificial Intelligence in Accounting: Practical Applications. Make sure to check that out, but just one more time, Cory and John, thank you so much for doing this with us today.
[Ng] Thank you for supporting us in this project and the opportunity for this podcast, the Pennsylvania CPA Journal for just the opportunity to write for them for all of these years and meet wonderful colleagues and think about these articles and all the great work that you do there. I just would also like to acknowledge Michael Colgan, CEO and executive director of the PICPA, who was gracious enough to write the foreword for the book.