 Hello and welcome to the session in which we will explore artificial intelligence in the profession of auditing. Auditors are exploring advanced technologies to enhance audit procedures and these days the focus on AI or artificial intelligence. Artificial intelligence can help computers learn, can help computers recognize patterns, can automate tasks. There's a lot of benefit for adopting artificial intelligence such as scalability means doing more work with less resources, reduction in cost. You can have savings when you can do more work with less resources which is scale your business as a result you can save. The information that you provide is more accurate when it's processed appropriately, enhance audit scope. You might be able to do more work if you are using technology and AI might give you some predictive insight that you may not see or human may not see or it may take more time for a human to recognize. Those are some of the benefits of adopting artificial intelligence. There are many more benefits which we'll talk about in a separate recording. Bear in mind AI comes with a lot of risks and we'll talk about those risks in a separate recordings as well. There are four potential artificial intelligence categories that can be used in audit. One of them is robotics process automation or simply known as bots. Two, machine learning. Three, deep learning. Four, natural language process, something like chat GPT. We're gonna cover each of these items separately starting with robotics process automation. Let's go ahead and get started. Before we proceed any further I have a public announcement about my company farhatlectures.com. Farhat accounting lectures is a supplemental educational tool that's gonna help you with your CPA exam preparation as well as your accounting courses. My CPA material is aligned with your CPA review course such as Becker, Roger, Wiley, Gleam, Miles. My accounting courses are aligned with your accounting courses broken down by chapter and topics. My resources consist of lectures, multiple choice questions, true false questions, as well as exercises. Go ahead start your free trial today. RPAs or Robotics Process Automation revolutionize the manufacturing environment and now they are moving into the routine office tasks. These technologies are known as bots. Now when you have a bot you don't have to have a physical person, a physical robot in that sense. A bot can be simply a software that can replicate assembly line functions for traditional business processes like data entry. So you don't have to have robots sitting there on the computer. The robot is inside the computer that's why it's known as a bot. Computers program with rules, executed tasks, they can do it quickly ensuring precision and efficiency in process based on established protocols. I can give you a simple example from my own company. I use QuickBooks online for my bookkeeping and recordkeeping. I can program the software to recognize when I have a revenue transaction, when I have an expense transaction in the bot and those rule-based tasks. They can quickly, not quickly, automatically, instantly recognize the entry, journalize the entry, send the information to the proper journal and the proper financial statements. So rather than hiring or doing it myself, the data entry, the system will do it for you. And this is a form of automation using artificial intelligence. RPAs involve deploying software, as I mentioned, to perform rule-based tasks, replicate in human interaction, like processing journal entries. Now the Big Four, they are using these systems and I'm going to show you a few examples from the Big Four, from the real world. Computer's terms as bot execute rule-based tasks swiftly, quickly, automating protocol with precision and efficiency. Oftentimes, when you go to a website, you first log in, you would see a chatbot. You'd say, okay, what, not a chatbot, a chatbox. And you think you're talking to a person, most likely you are talking to a bot. And this bot is programmed to answer questions from the website itself. So what happened is they trained this bot to learn everything on the website about the product, about the prices, about cancellation policies. And what you do, you ask a question. The bot will answer the question based on the information on the website. Now after maybe a few minutes, if you kept asking questions, the system knows that the bot is not being helpful. And as a result, a human might intervene and answer your question. But oftentimes, maybe 80 to 90% of the time, the bot can answer your question. Now these bots, they also might be known as intelligence agents in your CPA review course. You might see a different term or in your auditing course. That's fine. Now your clients, the clients that you're auditing as an auditor might be using those bots. And as a result, you have to be familiar on how these bots are being used because a bot can read sales data from emails or from forms, generate and send invoices, update customer accounts, manage overdue tasks and record issue resolutions. Rather than having a person, now the bot is doing it. Now, as business integrate those RPAs into their processes, auditor must assess the process efficacy as an integral audit component. Now what you're doing, you have to assess the controls, the risks of these bots as just you assess the internal control, the manual internal controls, the computer internal controls, you have to assess the internal control, the risks for using these bots because you're going to be relying on them as your client is relying on these bots. So auditors are increasingly embracing robotics in their practices. They have to audit bots, just like the customer will have bots, the auditor will have their own bots to streamline repetitive tasks like data transfer and reconciliation across system. So rather than an individual transferring the data from one computer system to the other or reconciling information across systems, the bots can do that. For example, an auditor might employ a bot to extract customer information from a client account receivable subledger, facilitating the creation and electronic mailing of confirmation for testing purposes. Who knows, maybe the bot now can do all of this. You have a program that goes into the account receivable subledger, select certain account based on the criteria that you told the bot. They will create that list, send confirmation to the customers electronically without human intervention. Now what does that mean? Well, it means you are eliminating the person from this process. So these bots can handle confirmation replies as well. So when the customer replies, the bots can look at the reply and compare the reply to the ledger to the subledger figure and reconcile the two numbers. If there's something wrong, if the numbers don't match, the bot might send another request or it might flag it for human intervention. The auditor use of robot, they can also use it for data extraction and data validation. For example, auditors can employ RPAs to extract financial data from diverse sources, validate the accuracy of the data and reconcile discrepancies. Now I know an individual, a person in the medical field, and that's exactly what she's doing now. She's programming an artificial intelligence bot to extract data from medical report and medical research for use for additional information for the medical field as well. But rather than going over those hundreds of pages of a medical report or a research, you would send the bot to extract this data, make sense of the data and provide the data to users. And I hope she will be successful. This not only reduce human errors, but also accelerate the overall audit process. Now ultimately, incorporating bots allow auditors to allocate more time to high value tasks. And this is one of the risks that we might consider is eliminating human intervention, which is it means, does this mean less jobs, not at all. But now, rather than performing manual repetitive tasks, you know, pulling information from a sub ledger, meeting a confirmation, that's not value added. If the computer can do it, you might use your time for some high value tasks. So don't worry, it should not replace your job as long as you adopt properly. So this is the use of bot machine learning is the second artificial intelligence technology we're going to be discussing today. Technological process enabled computers to identify patterns and create algorithm for driving solutions from data. So what does that mean? It means the computer is learning from itself. This technology allow computers to learn from data progressively over time to improve its performance without explicit programming. So what does that mean? It means the computer is programming itself based on these repetitive tasks and the data that it's using. And that's very powerful. Essentially, what's happening machine the machine, the computer is engaging in an inductive reasoning. Now, when we say machines, I don't want you to think of a physical computer, it can be a physical computer, but it could be just a software analyzing data to formulate predictive means it's given you some feedback about the information. So machine learning employs statistical method that enables computer to learn. That's why that's why it's called machine learning from the data gradually enhancing tasks performance. So as the machine learn more year after year, it's going to become much more efficient, better at handling tasks. So machine learning algorithm when applied to other procedures can rapidly quickly analyze a lot of data, vast data sets and identify anomaly, something that doesn't make any sense or trend that might otherwise remain concealed from human because human don't have the ability, the time, the resources to analyze this data. A machine can do it much, much more quicker. So for example, if we are performing a fraud detection assignment, machine learning models can be trained on historical instances of fraudulent activities and recognize similar pattern in the current financial transaction. So it can identify the trends, flag the transaction for you and you might be able to review it now investigate a little bit further. By doing so, the auditor can expedite the identification. It's going to tell you, look, these transaction might require further analysis and that's a great. So rather than spending three days trying to figure out which transaction I need to, I need to identify, well, I saved three days. Now I can spend more time on analyzing the irregularities, enable the auditor to focus more on investigative effort. Have you ever received a phone call from your bank or credit card agency inquiring about an unusual transaction on your account? For example, once I get a phone call, that why am I purchasing tennis equipment in London? Well, the bank recognize or the staff to recognize that I am not in London and someone when you was using my information in London to buy tennis equipment. Well, that's, that is an example of machine learning. The software at the bank notices that I don't buy tennis equipment and why am I buying this in London? It flagged the transaction. It did not process it. The bank automatically called me to verify it. Obviously I said, you know, it's either plus one, if you recognize the transaction, plus two to be rejected. Obviously I press two, then I stayed on the line and I talked to a person to let them know I am not in London. You know, if I need to cancel my credit card or my debit card, please let me know. And that was that. So if you receive a call like this, that's what's happening. The machine recognized an unusual transaction in the system. And this is what we may buy machine learning because the machine learned that's, that's not my habits. I'm not in London. I don't buy tennis equipment. So that's unusual. Flag it, stop it, inquire with the customer from the real world. This is an example from PWC, one of the big four, the software is called Halo. And this software, this software tests a huge volume of business, critical data analyzing whole population, improving risk assessment, analysis and testing, and unlocking a wealth of insight. So Halo is revolutionizing our audit by harnessing the power of data. So this is an example of AI that's being used by the big four. So you gotta get ready, you gotta know how to use this. This is an example from KPMG, another one of the big four. KPMG Research Credit Services with IBM Watson. IBM Watson, again, it's in form of AI, it's a, it's physically a robot, but it's basically a software. And what they're doing, they're using this technology to uncover more tax credit with less business disruption. Basically what they do, send this software to analyze long documents, where can we save taxes, uncover any, any business opportunity. Another example is Deloitte. Deloitte, they have a whole website showing their artificial intelligence. And you could go to those websites, for example, right here, artificial intelligence and analytics. And you could click and see what they are using. In other words, as a future auditor, as a future CPA candidate, as a future CPA, not candidate, as a future CPA, as an accounting student, you need to adopt, you need to be familiar. I will not be surprised if down the road, one of your courses will be about artificial intelligence or how to use artificial intelligence. Now, obviously, you know that the CPA is using more data analytics, more technology, I would not be surprised if down the road, we're gonna be having a whole course about technology within the accounting system, within the accounting information system. Machine learning and audit, as AI advances, auditor can leverage machine learning, as I just showed you with the big four, for stronger analysis. An example is, I just showed you Halo at KPMG, that's exactly what they said, they're gonna be doing. During analytical data auditing, ADAs, which we talked about, and in where entire population are audited, machine learning can absorb insights from auditors, handling of flagged exceptions. So ADA can test the whole data, because you're using the software. Now, these deep learning or machine learning can analyze those exceptions, if there's any exception, they can weather, you know, if you repeatedly accept or reject an exception, then the system that the machine can do it, repeated acceptance or rejection pattern enable the machine to identify and apply trends. So rather than an individual, accepting or rejecting, if there's something unusual, the computer can do, can perform the tasks, enhancing differentiation between regular and exceptional transaction within the ADA. So the ADA can analyze the transaction, machine learning, if there's any exception, they can either read it, accept or reject. Now, another example of machine learning is, you see it every day is the on YouTube and Netflix. Now, my son always asked me, he's seven years old, how do they know what I like? Well, the machine YouTube, basically YouTube will have some sort of a machine learning analyzing your views, analyzing your likes. So when you like a video, YouTube, the machine learning at YouTube would know, for example, you like this, for example, my son, what my son, Adam watches MrBeast and he likes their videos and he comment on them. As my son, watch and likes those videos, what's going to happen? YouTube would recommend more videos about MrBeast. If MrBeast has a new video, YouTube would recommend this or they would recommend similar channels to MrBeast. Same thing with Netflix. When you watch something on Netflix, it knows what you like, it knows your age, then it would recommend based on machine learning, based on what other people are watching, your age, your population group, and make the same recommendation. This is a form of machine learning. So you just kind of get, kind of just want you to get your head wrapped around this. It's, you're using it every day without even knowing. That's what I'm trying to say. So make sure to adopt, accept, leverage to your own advantage. Now, let's talk about deep learning. Deep learning involves training computers to recognize extensive data here beyond human capacity. With enhanced computer power and data storage, machines now can mimic human brain processes to handle vast data volume. Through repetitive exposure, the computers refined their understanding of anomalies to minimize errors. A classic example of this is the self-driving technology, which employs deep learning. To navigate the road, you have to worry about so many different things. The cars on the road, the lanes, the speed, the people in front of you, the people behind you, the people on your right, the people on your left. The computer can read all this data quickly through deep learning and react. Deep learning is being explored also in the accounting field. Now how? Well, well, you know, it's use your imagination. This dynamic blend human expertise and machine capabilities in shaping the future of auditing here. You got to use your imagination, how we can, we're going to be used deep learning and auditing and in accounting as well. Natural language processing. Prime example of it is chat GPT. Another avenue of AI integration in the audit is AI, is artificial intelligence chat GPT. This technology enabled computers to understand and process human language. And this will facilitate the analysis of taxual data. So the auditors can leverage this NLP to analyze contract. So rather than reading the contract, maybe it's 15 pages, you can upload the contract and you can ask questions about the contract, where is the risk, where's the risk in this contract. And the NLP will analyze this data or you could embed a financial report provided to you by management and ask questions about that report or other written document. And this is going to save you time. And the NLP will not miss any data. Why? Because it's reading every single word. For instance, NLP power tool can swiftly sift through organization contract to identify clauses that might pause potential financial risks or you can program it to look for that specifically. So this capability streamlined the review process and ensure that auditors can concentrate on critical areas requiring their expertise. Once the risk is identified, then you can go in there, talk to the client and give them some advice on how to handle this financial risk. What we're going to do, we're going to look at another recording to discuss the risks and benefit of artificial intelligence. What should you do now? Go to FARHAP lectures, look at additional resources, multiple choice through false additional resources that's going to help you whether you are a CPA exam candidate or an accounting student. Good luck, study hard, adopt artificial intelligence to your advantage and stay safe.