skip to content

AI in the Finance Function

You almost have these superpowers of somebody else or being able to collaborate with somebody else that's not an individual.

Narrator:

Welcome to Accounting for the Future, A BDO Canada podcast. For financial leaders to navigate change and achieve business growth. We'll uncover the challenges financial leaders may not have dealt with yesterday, but will definitely have to manage for the future.

Anne-Marie Henson:

Hello and welcome to BDO Canada's Accounting for the Future. I'm Anne-Marie Henson and I have the pleasure today of welcoming David McKellar, who's a partner in our technology consulting service line, as well as a member of our Innovation and Change team. The Innovation and Change team is dedicated to finding new ways to improve our business practices and drive growth through the use of innovation and technology. David, welcome to Accounting for the Future. I'm really happy that you're here today.

David McKellar:

Thank you. I'm excited and honored to be here.

Anne-Marie Henson:

Awesome. I want to talk to you today about a very relevant topic. I think not a day goes by anymore that I don't hear about AI or ChatGPT. And I think a lot of us understand that in some way, AI and how businesses will be using it in the future is completely accepted. But I don't know if people fully understand or maybe some of us are still struggling to understand exactly how it's going to apply to our business or what we do on a day-to-day basis, and how they can maximize the benefits of AI. So, I'm hoping that you're going to be here to demystify that for all of us today.

David McKellar:

I will definitely do my best. I'll do my best. I look forward to that part of the discussion for sure.

Anne-Marie Henson:

Exactly. This podcast is obviously called Accounting for the Future. We want to talk about topics that we see are really relevant today and how they're going to impact the future of our business, but specifically as it relates to the accounting and finance function. So, what we want to talk about is not just artificial intelligence or AI, but how it impacts financial statements, financial reporting, CFOs and controllers who work in the finance function, and governance around all of that stuff. So, before we start though, I wanted to level set and not make assumptions that everybody understands AI and generative AI. So, why don't we just start there with simple definition. What is artificial intelligence? And what is generative AI? What's the difference between the two?

David McKellar:

For sure. No, and maybe I'll say this, I'm going to come at it from a less technical slant and really more trying to apply it to business. My background is really more in let's say, process engineering, operations management, as well as the financial side of ERP applications. So, I will definitely try to explain it more in the way that you, from an accountant's perspective, would get it. So, I think really simply, if to look at AI, the way I would define it is, it's really a development of computer system or systems that have some flavor of human-like intelligence. And there's variations of that. I think AI in that term has been around for a long time, but we've also seen very different increments of what it is and what it's capable of over the last several decades.

And so, often I think people think of AI as automation, and that's certainly one part of it. We think of things like deep learning and leveraging that as a way to learn about different things versus coding in terms of what that looks like. But simply, it's a prediction machine is sort the best way to think about it. And as far as how it works, I think quite simply the best way to look at it is that, hey, in this world that we live in, we have vast amounts of data. And AI is able to use that vast amounts of data in combination with math or algorithms to really identify patterns, make predictions, or business decisions.

Anne-Marie Henson:

That's great.

David McKellar:

Is that a good enough definition?

Anne-Marie Henson:

Yeah, absolutely. And when we talk about generative AI or gen A I, is that the same? Is it different? How does it play into the definition?

David McKellar:

For sure. Yeah, and I would say that it's almost like the next level. So, I think if simple AI is things like automation, generative AI has really been, we'll say the latest, greatest combination of all these technologies, where now we have this class of artificial intelligence that's actually able and has been designed to go and generate new content. And now that might be in the form of a text. It could be an output, like a picture or an image, but it's the fact that it's actually generating content based on all of this data that it's been trained on.

Anne-Marie Henson:

Okay. That makes a lot of sense. So, we're five minutes in and I already feel like I've learned so much more than I knew before. So, thank you very much.

David McKellar:

There we go. Yeah.

Anne-Marie Henson:

So, I do come from an audit and accounting background, so I want to start with that. But really more about how AI or gen AI can help me with my day-to-day job. So, a financial statement auditor, which is what I still do on a day-to-day basis with my clients is, our responsibility is to audit the financial statements. So, we start with the numbers that are provided by a client. We do a risk assessment. We analyze trends or variance from previous years. We get updates on what the client's done during the year. We assess risks, and then we'd assign procedures to address those risks. And then at the end of the day, we're able to sign off that the financial statements are reasonable in the context of the work that we did on a sample basis. So, in a nutshell, I've just explained to you what an audit would be and would entail. So, when we think about AI and the benefits it could provide to me to make my day-to-day job easier, how would you see that working? What could I use AI for at some point?

David McKellar:

Yeah, for sure. And I think AI, I think in the accounting realm has already been serving a role for a while in a few ways. I think one, there's an efficiency and an automation that AI can enable that otherwise couldn't be done by humans outside the fact that you would require many of them. So, if I think of something like for example, testing, I think that's the audit term testing, when you're looking for different patterns instead of a human sorting through different files, and there's only... No matter how fast they are, there's only so much that go through. Now we can actually leverage AI to actually drive through that much more data and still pick up on the trends. So, there's a level of efficiency, but there's also a level of quality that machines can work to.

And I would say that it's not a replacement, it's an augmentation. So, the important part is how is it able to sift through more data to identify things that could be an issue that we want to look into, and now the human can jump in or the accountant can jump in, and actually make a judgment call. So, I'd say that's probably a very simplistic way of looking at how it's been already used today.

The other one is, is that if we even fast-forward and think of maybe what's coming, and I would also make the argument that it's here today based on what I see in some of the Instagram or Facebook ads, this whole concept of generative AI, where it starts to get, I think, really fun is you almost have these superpowers of somebody else or being able to collaborate with somebody else that's not an individual, it's a gen AI system. So, that whole thing of a role-based gen AI, that's looking at the data, looking at the patterns, looking at the financial reports, and now being able to have a conversation on the things that maybe you should be looking at that you weren't paying attention to.

And a great example that I had somebody kind of state back to me, we were chatting about a doctor. And I think somebody, unfortunately, I think their child had broken their arm and we were chatting about what that could look like in this future world. And they said, "Well, listen, instead of just going to one doctor to read the image or the X-ray, wouldn't you rather the 10,000 best doctors across the globe?" And so, that's the example. You could literally have this superpower of that would be this almost accountant-like agent or gen AI agent that's sifting through all that and pointing you to things that you otherwise wouldn't look up. Right now, that's probably a little bit more futuristic, but the reality is that stuff is being looked at today, right?

Anne-Marie Henson:

Yeah, absolutely. And I guess what's interesting is that the pace of change that's happened even over the past few years. I know the concept of AI and automation has existed for a number of years. But what's happened in the past year alone would probably indicate to you that what you're talking about, which sounds really interesting, isn't that far away. That might not be 10 years from now. Could be five.

David McKellar:

For sure. Yeah. No, and I mean I think what happened back in November 1st, 2022, I mean ChatGPT or OpenAI lit the match with the drop of that. And the pace of change since then, we've had several different releases that have come out with a lot more powerful features and functions. Microsoft obviously released or has released as of November 1st this year, copilot. So, these capabilities are getting integrated just so much into all of what we do. And I mean, AI has been integrated in a lot of what we do today. I mean, if you have, whether it's Amazon Prime and understanding, hey, what should you look to buy tomorrow? We all have music playlists of some sort, or maybe a Netflix recommendation. There's so many different use cases for AI that we've seen already in action. But gen AI has certainly stepped that up, right?

Anne-Marie Henson:

Yeah, exactly. No, that's really great. And I know I'm personally looking forward to it. Like you said, it doesn't replace the auditor. It doesn't replace judgment, at least not yet. So, it's a matter of finding a way to use those tools to make our job easier to detect things that you wouldn't necessarily have the time or the ability to detect on your own. But with this power, it gives you the ability to be more efficient or look at things differently than you would have, right?

David McKellar:

Yeah, for sure. And I think, listen, for the first time ever, the knowledge worker feels like under attack because of I think what we're seeing in gen AI. And there's certainly some of that. But I definitely am an optimist. And I think of the statement that AI's not going to replace people's jobs, it's going to replace the jobs of those that don't embrace it. And if we look at accounting, I mean, we went from a paper ledger to, gosh, everything's now on computer, or Excel, or some program. Can you imagine going backwards and actually saying, "Well, no, I'm not going to accept computers. I'm going to stick with my abacus in my journal." There's just a ledger. There's just no way. And we laugh about that because that is so backwards, but that's really the leap that we're really underway making right now.

Anne-Marie Henson:

Yeah, no, I like that. And I'm a fan of the optimistic view on things because otherwise, we'll get left behind by all this technology. Well, that's really helpful. And I guess I want to turn the page a little bit. Sticking with the finance function that may be in industry or at companies. So, I know a lot of our listeners are CFOs, or controllers, or analysts at companies themselves. And I have heard a lot about how this could help their job.

So, a lot of CFOs in the finance function are responsible for a variety of different things around financial reporting. They deal with operations, and they help to assess and analyze profitability. They have to coordinate with banks, and lenders, and investors to keep up to date on their lending requirements, but also often to go get additional funds when they need to. And they also help support a company's growth objectives, whether it be expansion or acquisition. So, let's talk about that a little bit. What have you seen? Provide us a lens into the future as you see it with AI. How could these types of tools help a CFO with what they have to do in their company?

David McKellar:

Yeah, for sure. And I mean, I think a few that I can think of that come to mind. So, I think whether it's efficiency or automation, I think there's definitely a streamlining of processes that AI can absolutely enable. And there's so many different reasons for that, whether it's compute power, whether it's Moore's Law, and just the fact that data is so much more readily available, and just that compute power and how much it's doubling up. So, I think really making and streamlining processes is definitely out there. And I think we've seen some of that over the last several years. I think the other part though is really pattern recognition and prediction. And just rather than having to do that manually or because someone's got a really great intuition, we're really able to leverage data, both the data that we have, but also potentially combining that with market data to make, I would think, much stronger predictions today than we really ever have in the past.

And then when you combine that with some of the things that we see in terms of gen AI, that opens up a tremendous amount of use cases. So, a great example as somebody was describing to me was twofold. One, they were walking through how in the future, how if you think of going and doing a stockroom basically count, just taking a simple picture. And from that, being able to do a count, that's a very great way to be able to streamline a process. The other one is just even from a safety requirement perspective, where literally being able to record and seeing what's going on, whether that's in a warehouse, or maybe it's at a site, or something like that, feeding in that video. And then in the future, being able to actually understand and train people based on what they're doing or not doing. I mean, that's more of probably a futuristic one out there, but I think coming fast.

The other part though is there's a quality level to this as well. So, I think just accuracy improvement. And I think we've seen that with computers. No matter how good we are at math, the reality is a calculator and a computer does that better. And now with AI and just what it can do in terms of enabling and automating processes, it just drives accuracy and quality that much better.

Anne-Marie Henson:

Right. Absolutely. No, it reminds me a bit of something I've heard a lot in terms of companies that are out there looking for funding. Oftentimes, these require you to do a future cashflow analysis. What do you expect? What are you going to do with the money? What do you expect to generate? What your profitability is going to be? And it's a poorly kept secret that those future cashflows never accurately reflect what really happens in the future. And then you have to spend all that much more time after reconciling why you didn't meet the cash flows, and going back to your stakeholders and your lenders to explain why there was such a discrepancy. So, that could be something really interesting as well. You mentioned the accuracy of information and being able to use the predictive analytics to be able to really forecast with a little bit more precision than we have as humans.

David McKellar:

For sure. No, and there's lots of great, whether it's cashflow forecasting, sales or demand planning. And some of those use cases, you see those coupled with ERP, or financial, or supply chain solutions. I mean, we've seen that for a while, which is great. And they're only getting better. I think that's the neat part with AI, is that it's really all about learning. So, the more that it gets out there and the better the models get, and the more that they get fed, the more they learn. And so, thus that prediction model is just getting better and thus the quality output of the predictions is just better and better.

Anne-Marie Henson:

I love that. That's really exciting to see how that's going to work for us in the future. And I did speak to one client recently. I thought this was an interesting use case to share. They were working on developing an AI tool that they were adding onto their system, where it was in their legal department. But I think they easily apply to the finance function as well, because legal reads, and rights, and signs the agreements, but there's always an impact on how to account for those agreements afterwards. So, there's, on average, this company signs about 6,000 different types of legal agreements a month. And every month there are changes to those agreements that aren't necessarily blacklined or highlighted for the department to read.

So, they were working on a tool that was going to help them collate these agreements and have the AI tool analyze them to give the team just highlights in terms of what the changes were instead of them having to read through those. So, I know sometimes technology can seem a little bit scary or that it's going to take our jobs. But I can imagine how much more this department can do with their time without having to comb through papers, pages of lines of different changes in legal agreements.

David McKellar:

Let's be honest, computers are going to be so much better at comparing two documents. And then when you think about the fact that how much the volume that you just said, was it 6,000 per day?

Anne-Marie Henson:

Per month. Per month.

David McKeller:

Per month?

Anne-Marie Henson:

Yeah.

David McKellar:

And I'm sure they're not single pages as well. And I think just the ability now with also gen AI has really strengthened that capability of also just being able to interact with it, summarize. So, that's such a great use case that you've highlighted there for sure.

Anne-Marie Henson:

Yeah, exactly. So, I did want to talk about one more aspect of this or a potential user of AI, or at least a group that needs to understand how it could be used and the risks involved potentially, and that's at the governance level of companies. So, talking about audit committees and board of directors, and what they need to be mindful of as this technology continues to grow and likely gets implemented in the companies that they're overseeing.

So, David, you and I just a few weeks ago, had the chance to attend an audit committee round table where we hosted about a dozen different audit committee chairs or board members of a variety of different companies from different backgrounds and industries. And it was so interesting, really such a great rich conversation. But I did want to bring it up here and give a bit of a platform for us to talk about it a bit more. So, I'd like to hear your thoughts. If you were going to give some advice to someone who is on an audit committee or on a board of a company that's looking to adopt AI and use it in their company to help with their growth, or efficiencies, and things like that, what are some of the things that they need to be mindful of?

David McKellar:

For sure. Yeah, and I think there's a few things. I'm going to go to the usual suspects in a second. But I think the one thing that for me struck a chord coming out of that conversation was just how of those 12 people, everybody was at a very different level in their journey of understanding what AI is and how they could truly leverage it within their company. And I think it's really not if, it's really a matter of when. So, if you as a financial leader are looking at it saying, "Hey, should we do this?" The reality is absolutely, would you ever not use computers? Would you ever not use the internet? I think the first part of that is really, truly education. I think be curious, learn as much as you can about it. Yes, absolutely, there's a safety element to it. But I would really employ or encourage everybody to learn as much as they possibly can about that.

And I think some of the other usual suspects in terms of just governance is a big one. And governance is a few fold. One, I think it's the data side of things and just what needs to be locked down, what could that look like? The better your data is, the better the models are going to figure out in AI in terms of how it can be leveraged. So, just understanding that whole data management cycle. I think that there's also a level of just extra level of security, and privacy, and governance that has to go along with that as well. And we've seen this a little bit play out in that often documents are secured or data secured because of truly obscurity. People don't know that it's there. They don't think to ask for it. But all of a sudden now with this superpower of AI or gen AI, you can find that data pretty darn quickly.

And it's been a little telling to find at times, just maybe some of the holes that you just didn't understand as far as managing and controlling the data. So, I would say that security aspect is one, and just that full governance. A couple other ones too, I think is transparency. I've dialed a phone number, and we're talking to a bot, and it drives us nuts. Well, that is less and less becoming the case. At times, we're talking to a virtual agent and we don't even know it. But that transparency of the fact that we're talking to AI or something's been generated by AI is key.

And it does go hand in hand, I think with just the ethical use of AI as well as responsible build out or how you build AI solutions for your organization. That is a key part. And I think the cool part about this is that really it's been this industry in terms of accounting, if I think of whether it's been SOX or things that are going on in terms of ESG, you guys have really been on the forefront of that. And so, I think we're going to see a level and a layer of regulation come into being that will likely create some opportunity and some challenges from a finance function of just understanding what's there. Are your algorithms, are they biased in any way? Are you managing for that? Again, there's some really neat things to think of from that perspective.

Anne-Marie Henson:

Yeah, for sure. No, that's really great. And I agree that in terms of the diversity in that room where we did the round table in terms of where people were in their journey, that everyone really understood that it was going to be something that was going to form part of what they had to be mindful of and care about in the future. One of the things that struck me that I don't know if I necessarily thought of before, but I think is important as well. And you did talk about it a little bit in terms of quality of the data that comes in, even to add to that is the potential bias in that data.

On one hand, you've heard AI is great because it's not a person, so it doesn't come with systemic biases or it doesn't necessarily assume that it has the right answer. It only thinks about things in a specific way. But if the data that's been fed into that model or that machine comes from a place that has only looked at things a specific way or lacked diversity in any way, whether it be a diversity of thought, it could lead to giving you the wrong answer or just not giving you the full answer. So, I thought that was something that was really interesting that I think from a governance perspective is really important to challenge to understand where does that data come from? How is it fed into the machine? And how can we ensure that it's considering all potential options and not just the ones we think were right?

David McKellar:

For sure. And I think the more that AI starts to enable an organization, the more important... And it's always going to be important, but the more important that understanding of how biased is your data and your algorithms, because you're going to be asked for that. Just like today, I think many organizations, whether it's maybe a customer, a vendor, a lending partner, part of the question they're asking is DEI or ESG. And I can absolutely see a time where they're going to ask for how biased our data or algorithms are, and can we prove it?

So, I think there was a statement that I had heard, I'm trying to remember. It was a gentleman from Microsoft that had said it. And it's a really good case for diversity. And the way he said it was, AI is going to represent the world. But for it to do that, it needs to be representative of the world. And it does speak to that. And just constantly challenging ourselves in terms of, hey, what is that level of bias? And is there, and what is it, and how do I fix it and manage for it again in my data or in my algorithms?

Anne-Marie Henson:

Yeah, exactly. So, I have one last question for you, and it's not entirely related to the finance function, but you do seem to embrace this technology a lot. So, I'm going to ask you your own personal thoughts on something. We've read so much about ChatGPT, and OpenAI, and what's been happening in the news in the past year. And there are some individuals, you would call them the founders of the AI movement, who've come out publicly and have shared some fears and anxiety about what AI could do in a negative way. So, I'd like to hear from you what you think about that. Is this something we need to be mindful of? Are you optimistic about the future and what AI can do for it? And really, where do you see this heading?

David McKellar:

No, and good question. And I would say 95% of the time, I'm an optimist in this. Every once in a while I might have a bit more of a non-optimistic slant. I think that, I guess a couple things. One, the genie's out of the bottle. So, I don't think our ability to say, "Hey, you know what? Just kidding, let's go back." That's not going to happen. So, do we need to be having the conversation about ethical and responsible AI? Absolutely. The more people that includes, the better. That being said, I also do recognize that there's going to be bad actors in anything. And that's why it's just that much more important for all of us to truly get educated and embrace it and understand how to use it and drive it, and build use cases that really do drive humanity.

And I think what's also been interesting is that you've seen some of the people that have had the cautionary tale that I would say are more and more understanding that, hey, they have to embrace it or at least be part of the conversation. And so, I think that's really what it is. It's an education. It's about being part of that conversation. And how do we make sure that we're employing what is going to be amazing, and I would say society changing technology for the greater good, right?

Anne-Marie Henson:

Right. Okay. Well, I like that answer. It makes me feel good about the future. And I agree. I think at the end of the day, it's up to the majority of the population, those who want to use this technology for good and not for bad to make sure that we hold ourselves accountable and stay, I guess aware of what could happen. But I do like the positive outlook.

David McKellar:

And it's funny, I think the one parallel that I've read a little bit about that I think is actually a really interesting parallel, it was the advent of the printing press. And I'm trying to remember which it was king when at the time, where this... What was really you think revolutionary technology outlawed the printing press because it was creating an ability for more people to consume and understand information versus just what was either written down in a book or from word of mouth. Now, the funny part was though, that was also a huge change in terms of society where all of a sudden people could learn so much more and really that democratization of learning. And that's really, I think what this becomes about, is that this technology does truly, when it's leveraged and used well and by everyone, it opens up so many different things for really everybody. And so, that's the part that does excite me, is just the democratization of knowledge is now at our fingertips more so than it's ever been.

Anne-Marie Henson:

Yeah, I really like that. And thank you for your comments and our conversation's been wonderful. And I have a feeling that if we talk this time next year, maybe there'll be even more stuff to talk about just given how fast this is changing. So, really, thank you for your valuable time and input today. I hope our audience appreciated this discussion. I'd also like to thank you, our listeners, for tuning in and to tuning in to all of our episodes. I'm Anne-Marie Henson, and this has been BDO's Accounting for the Future. Please let us know if you found the topic interesting and useful, and remember to subscribe if you liked it. We'll see you next time.

Narrator:

Thank you for listening to BDO Canada's Accounting for the Future. Past episodes and related insights are available at www.bdo.ca/accountingforthefuture. Or you can go to Apple Podcasts, Spotify, or Google Podcasts to subscribe. For more information on BDO Canada, visit bdo.ca.

This site uses cookies to provide you with a more responsive and personalised service. By using this site you agree to our use of cookies. Please read our privacy statement for more information on the cookies we use and how to delete or block them.

Accept and close