The CFO and Generative AI – A Transformative Partnership

Neil Morrison:

billion in: 2022

Over the course of five seasons on this podcast, we have explored a lot of the forces that are driving change in the CPA profession, but to be honest, nothing we have covered feels as significant as the rise of this new technology. That's why this season we are taking a deep dive into the impact of Generative AI on the role of CPAs and on the finance function in general, and we're going to start at the top with the CFO.

The role of the Chief financial Officer is undergoing a significant transformation thanks to the advent of Generative AI, and two people who have had a front row seat to this transformation are Sree Balakrishnan and Zohaib Akhtar. Sree is a partner at Deloitte who focuses on technology and data specifically for finance organizations, and Zohaib is a senior manager in Deloitte's Omnia AI practice. Together they run this lab where they expose finance professionals, including CFOs, to the possibilities of Generative AI. They run them through exercises to show them what works, what doesn't work, the opportunities, the pitfalls, and when the session is over, Zohaib says the reaction is frequently one of disbelief.

Zohaib Akhtar:

We've seen many a times of folks who come and join these labs be stunned by the types of examples of applications of technology that they will come across during the day's activity, so that's one, obviously looking at what the possibilities are. What I've also seen is looking at requirements. One is looking at, yes, this is what is possible, but then also when they look at the requirements, oftentimes, it really truly surprises them how simple certain situations can be.

Neil Morrison:

Sree says the sessions can also help calm some of the fears about Gen AI.

Sree Balakrishnan:

I think one of the things that they realize quickly as we go through the use cases is that it is not Gen AI replacing humans, it's more giving you a headstart. Rather than sitting down with a blank piece of paper and creating a narrative for your investors that you have to go and present to your analysts, it'll give you a starting point that then needs to be tailored by a human. Really, it's just removing that writer's block, if you will. When we use those kinds of terminologies, it's an aha moment for, it's like, "Yeah, okay, I can see now how I can use it. Yeah, I don't have to fully trust and blindly send it to the auditors or external folks," but it's really having somebody in between human involvement to validate but they're not starting from scratch.

Neil Morrison:

Deloitte recently surveyed CFOs on their adoption of AI. The goal was to see where they are in what you might call their AI journey.

Sree Balakrishnan:

I think, in general, the organizations have started, CFO specifically, have started to experiment or understand in two buckets, understand Gen AI and also starting to have planned to experiment using Gen AI in some way, shape or form. That's very promising and there is an urgent ... I guess there's a desire to get educated and see how it can be applied. Ninety-four percent of corporations, I think we interviewed about 2,600 plus organizations, and they said they are going to be looking at AI as a key success strategy for them in the next year.

Neil Morrison:

Sorry, what percent was that again?

Sree Balakrishnan:

Ninety-four percent.

Neil Morrison:

Wow.

Sree Balakrishnan:

It's a pretty big number, and we're seeing more and more of clients and organizations asking for, just in general, "Hey, what are some of the use cases? What is this technology about? Get my broader team across this new technology, what it can do." I think that awareness-building is certainly picking up.

Neil Morrison:

If 94% are thinking that this is going to be fundamental to how they move forward, essentially that's all of them. But they are still at the stage where it's exploratory at this moment. They're trying to figure out, what does that look like? Maybe Zohaib, from the standpoint of the CFO, right now, what are the biggest potential opportunities from the Gen AI evolution that we're seeing?

Zohaib Akhtar:

From a CFO perspective, think of it this way. A CFO would look at the business and would have an objective to increase as much, I would say, cost efficiency as much as possible. That would be, I would say, a key objective for them. Another piece would be to work with the other leaders, other senior leaders within the organization and focus on bringing about increased productivity because that would then have a direct impact to their revenues. Keeping those two things in mind, firstly, the bottom line from a cost standpoint, and then secondly, productivity, and then your top line, your revenues, your CFOs would be very keen on understanding how they could leverage AI and bring about those benefits, those improvements.

Especially in today's economic conditions that we see right now, the challenges that almost all organizations face, this is something that I would say is going to be key and critical for them in order to succeed bringing about advancements from a productivity standpoint, things like bringing in RPA, Generative AI, bringing in efficiencies, bringing in education, getting their workforce updated. All of those things would be, I would say, front and center for them at this point in time and should translate into hopefully larger savings and better revenues.

Neil Morrison:

That seems pretty spot on for what the CFO wants to happen. Is there an example you can give of that, of how Gen AI in particular can be used to do that? Just either reduce your costs or to increase your revenues?

Zohaib Akhtar:

Sure. Think of it this way, and I'll probably layer in something very, I would say, interesting to CFOs again. It is the topic of ESG. When you look at the environmental sustainability guidelines and the fact that this standard is now going to be coming into play as the world works through that challenge of climate change, what happens is, again, this CFO finds themself on the hot seat. What they can do with Gen AI is they can look at finding opportunities within the organizations to reduce and be, I would say, carbon-friendly if I can. Essentially, what they would do is they would use Generative AI to understand the regulations, make it more, I would say, reachable across the organization, bring about more understanding. Then, they could use AI, so we'll take two steps back and look at AI as a whole and see what sort of abatement activities can be brought in.

If I'm an organization that has a lot of spend in terms of transportation costs where my fleet of trucks, for example, does deliveries and is using a lot of gas, I probably want to have abatement bring about ways where I can reduce the amount of gas that the fleet is using. That could leverage, you could say, technologies involving AI, that could bring about predictions and I would say guidance in reference to what the organization needs to do to hit those targets, and/or leverage things like IOT, the internet of things, have devices across multiple trucks or within their fleet and get the necessary measurements within time and basically make those type of things happen. I would say, again, that could be a classic example of using Generative AI, AI, IOT, all sorts of technology to essentially hit their ESG targets.

Sree Balakrishnan:

I think the other, from a cost standpoint, cost savings, CFOs and financial organizations in general, there's multiple use cases. Even looking at your closed process, for instance, as you go to month-end or quarter-end, the finance teams spend quite a bit of time and effort in getting the books closed and then having the reports generated for internal decision-making. Gen AI can really help with those data entry, manual data entry or manual journals being uploaded into your ERP, helping with identifying variances earlier in terms of what's in the operational data that you're collecting to what's going into your lake and into the ERP, identifying those variances and even get to a point where it can start to recommend or highlight recommendation issues and also give some recommendations in terms of next steps based on how you have resolved prior variances. Those are simple examples of where GEN AI can start to really improve the operational efficiencies and thereby start to help with some of the cost savings as well.

Neil Morrison:

That's the plus side of it. What are the biggest potential pitfalls? If I'm a CFO, what are the things that I need to be cautious around?

Sree Balakrishnan:

A couple of things. One of the things you want to look at is there's going to be changing regulations coming up. There aren't a lot of guidelines yet in terms of how the data is going to be used. You can hear it in the news that there are specific technologies, especially Gen AI, that are getting sued now. I think the most recent one was New York Times, it's all copyright-related. It's really looking at how are you using your customer data. If it's private search functionality or private Gen AI, then again, you have to look at how you are using customer data, PI data in your models. Some regulations like that will come up and how you evolve and adjust to that.

The other one is having human in between. Gen AI still has issues like hallucinations, where it could just create content that may not make any sense or may not be accurate and it's a real issue. Being able to identify that early when these drafts get created, so again, goes back to not blindly taking what the technology produces and then giving it to external audiences or internal stakeholders but having somebody in between to be able to proofread and review it. I think that'll still be very important. The other one is, just in general, having the right skillset. How do you have the right team in place to really help to navigate this technology and the evolution of the technology across use cases?

Neil Morrison:

Yeah, I'm glad you brought up skills because I'm wondering what sort of skills should CFOs look to develop within their teams to effectively use Gen AI?

Sree Balakrishnan:

Yeah, I think, and Zohaib, you might have some views on this as well, it's really, you're going to have your CPAs and your traditional finance folks who are going to be the knowledge experts that you need to have on the ground. You also need folks now who are going to be familiar with what the underlying data is, how to make effective queries, and then looking and having understanding of how this technology works so you can actually then formulate different use cases and where you can actually help get the most out of it.

It's really, in the beginning, it's almost having a couple of SMEs in your team who understands the technology and are able to educate the broader group and look at how you can utilize it, but going forward, it's going to evolve. People who are more familiar with technology and can bridge that technology and finance angles, it will be important. I look at it similar to how we went from, what do you call it, manual bookkeeping to Excel and a change that may be there now Excel to Gen AI, it's a similar sort of shift. There'll be more additional skillsets that'd be needed and different things will come up, but at the beginning, at least having SMEs who understand the technology to help the finance teams navigate, it'll be important.

Neil Morrison:

AI experts, making sure you have the CPAs, you have the finance people on board as part of the team, but you need the AI expert in the mix.

Sree Balakrishnan:

In the mix, absolutely. Absolutely, yeah.

Neil Morrison:

Zohaib, if I'm a CPA, what skill should I be developing to ensure that I'm able to capitalize on the growth of Gen AI in finance?

Zohaib Akhtar:

I'd say three things or three skills rather that would very much help you. Again, being a CPA, you'd be good with numbers. Generally, when you're good with numbers, you're also good with working with tables and tables would also translate in you becoming or rather being good with data, because data is all structured or based out tables generally. I would say, being proficient with data, tables, numbers would be number one. The second piece would be being really good with their statistics side. Machine learning as an example, it's all statistics models. As long as you've got a decent understanding of, I would say, the engine, what's in the hood, it is really helpful as a CPA as you get different models applied.

Lastly, being creative. This is now a soft skill. Being able to think out of the box and being able to, I'll take the classic example. When accountants started working with Excel, Excel was just a spreadsheet, it was a spreadsheet tool. Over the years, we really evolved into using Excel for so many different things. The formulas came in, and the graphing, and the Pivot tables and so forth. It evolved into something very different than what it was really envisioned back in the mid-80s when it started with Lotus 1-2-3. It's the same idea here, that if they are creative, if they think out of the box, there's so much more that they can do.

Sree Balakrishnan:

I think that's a very important piece that you hit. I think from finance perspective, I think usually, typically you looked at as a guardian, especially in the controllership, that's still very, very important, that trait is extremely important. But being open to understanding this technology and being creative, being inquisitive, and don't be afraid to use this technology and apply to everyday, your roles, I think that's very important, because you need to evolve with it and there will be new use cases that you'll come up with, and so having that mindset of curiosity and being inquisitive and using the technology, being open to it, it'll be very important.

Neil Morrison:

There's a real tension there though, isn't there? Between being the guardian, the one who's trying to keep everything, being the controller of all of this, and then also, "Hey, everybody, let's take some risks."

Sree Balakrishnan:

That's where it is a fine balance, but I think that's where you look at those specific use cases. You start with something really tangible, really small that you can realize an outcome fairly quickly. I think that's very important. How do you identify those specific things, even if it's just creating drafts? Drafts of reports, drafts of a narrative that you can then review. It just makes your life that much easier and more productive.

Neil Morrison:

How should CFOs measure the success and the ROI of Gen AI when it's implemented in their finance functions? How can they measure whether this is successful or not?

Zohaib Akhtar:

Think of it this way. From a CFO perspective, you've got two buckets. You've got the bucket of accounting, you've got the bucket of finance. The difference between the two could be on the accounting side. Things like your month-end close process, things like processing your journal entries, going through your executive reporting. A lot of these things could be automated and from a CFO perspective, if I'm quantifying or if I'm measuring success, let's say the number of hours that it used to take, the amount of time it took to, let's say, get through those reports has now been significantly reduced, and that could be your measure of success.

On the finance side now, what happens is making investment decisions. As you look at different types of scenario analysis, all of those things could now be simplified and you could have fairly good guidance provided from these Generative AI or AI tools that will support decision-making and will allow you to make decisions that will ideally maybe let's say help you increase 5% in revenues. Now, that additional 5% is your measure of success. Essentially, that's how you would look at it.

Another point I'll also add here, which is probably very important, is that you have to do two things. One, you have to be surgical in your approach. Pick out specific use cases and go after them. The second thing you must do is you must ensure that you're not going to aim for perfection. Now, that's something that the finance folks will often get stuck in where they want, let's say, a two plus two equals four. What I would recommend in this instance is that perfection would be the enemy of progress here. As long as you're directionally right, then maybe that's the decision you make and that's how you make progress.

Neil Morrison:

You guys have both talked about this very targeted approach that's required, what's the right way to identify those targeted opportunities? How do you see the things that are obvious targets for you to experiment with or apply it to?

Zohaib Akhtar:

I'll give you an example and I'll probably start off with an analogy. If you look at autonomous driving, as a whole, what tends to happen is it's not probably quite there where we probably would've wanted it to be. Meaning, if I sat in a car that drove on its own, I may still be a little tense, a little anxious, maybe not even, I would say completely comfortable. However, if you now break down autonomous driving and look at things like safe lane changing or keeping distance from the car ahead of you, most of the new cars now have those built-in features, and I'd quite confidently say that those features are quite functional. I think we feel comfortable using them, we trust them and they do a really impressive job.

I would probably look at it that way, that as you look at the entire process, let's say for a month-end, you could have a targeted approach to see how you could, for example, reduce the month-end close time. Or when you do your executive reporting and you have your commentary that is completed, so there's an MDNA team often, let's take the example of FSIS. They have quite a large team that works through that process, takes long hours and gets that information ready. That sort of stuff could now be targeted and taken down as a use case, and then we could bring in really good strong results and that could be, I would say, a strong win.

Sree Balakrishnan:

Yeah, I would just add to that. I think you want to start with something where you have the data collected as well. One of the things that you need to have is a data in order in terms of a clean data set that you can work off of, because otherwise, it is going to be garbage in, garbage out. If you have use cases where you have a certain set of data already there that you can make use of, then those are the ones that you want to probably prioritize and see if we can get some gains.

The other one is, I think, we want to keep it pragmatic. There's a lot of hype out there in terms of Gen AI as well that it's going to solve everything, no need for humans and it's going to get cost reduction to a great extent. That's not there yet, it's not really the case. When you look at these pilots and start with something so small, you're looking at 15%, 20%, 25% improvements in terms of efficiencies, in terms of what you can do in productivity improvements. Do you want to keep that level ahead? It's not going to get rid of your process altogether, and so you want to just approach it that way and those are the gains that you're going to get.

Neil Morrison:

With that caution in mind, I'm going to ask you to throw it aside and look into the future. What are some future developments in Gen AI that you can see coming down the road or potential that could significantly impact the CFO role?

Zohaib Akhtar:

I would say, one of the first things you're going to see with Gen AI is going to be the democratization of data of insights. The speed that CFOs have access to information is going to significantly increase. It's almost going to be real-time. I think that's definitely going to be a gamechanger because now they're going to have all the stats, the live information they need to make decisions, which will naturally bring about lots of changes, improvements, and definitely bring about some dramatics as well.

In addition, I would say the simplicity. When I say simplicity, I'm talking about your day-to-day mundane tasks probably being completely replaced. You may also see Generative AI, for example, maybe even taking out your prompting. Today, prompt engineering or prompting is quite a key component of this angle of technology, but what may also happen in the near future is that prompting would potentially completely go away and it would be embedded in our day-to-day tasks, and that would again make our processes quite powerful and simplified as well.

Neil Morrison:

I like the image of the CFO sitting in the meeting. Someone asks a very difficult question of them. They no longer have to say, "Yeah, let me talk to my team and get back to you on that." They just type it in and get the answer straight away.

Sree Balakrishnan:

Yeah. No, that's going to be super powerful, and it's also leading on from there, it's also real-time simulations. Being able to create, we call digital twins, but really running simulations on different variables and if you're reporting on a number of KPIs every month or every quarter, what are impacting those and how can you adjust levers around those KPIs and gain that information early often in real time? I think that's going to be fundamental.

The other areas, I think there's going to be a lot of advancements on incorporating Gen AI. The vendors are really aggressively implementing tools within their stacks, and that's going to have an impact again. There's a lot more real, I guess, provisioning of reports and being able to understand insights, as opposed to in the past there's been promises of finance being able to do things themselves but really they always needed a technology team to make the actual change or get the reports. This is going to really help change that narrative where Gen AI can make data more accessible, more finance users to be able to understand what's there and be able to query it in real-time and get information out of it and build their own insights. I think that's something that's going to change that function quite a bit and how they work with IT.

Neil Morrison:

Zohaib, you had a last thought there?

Zohaib Akhtar:

Yes, I did. Adding on what Sree is saying, I also strongly feel that the CFO themselves, they'll find themselves playing a greater role within the organizations. I may go ahead and kindly say something a little radical here, but I will, what we often find is if you pick up a chief data officer, you pick up a chief information officer and you ask them to wear a different hat, ask them to wear a hat of a CFO. They may struggle potentially, but if you ask a CFO oftentimes, to wear the hat of a chief data officer or even now, I would say in the future now with the advent of Gen AI and all the technology advancements, the expectation would be for these CFOs to play a larger, a greater role.

I see them playing, I would say, a strong role within those other two areas as well. Meaning, a CFO could potentially also in combination be a chief data officer and/or a chief information officer bringing their strong background, their understanding of the business, and also having support from AI to understand, again, how technology is functioning within the organization and then potentially moving forward in that direction.

Neil Morrison:

Oh, that's fascinating. What a great conversation? This has been so helpful. I really feel like you guys have really laid out in a nice concrete way how this is impacting the CFO role and how it could potentially impact the CFO role, so I really appreciate it. Thank you guys so much.

Zohaib Akhtar:

Thank you. Thank you for the opportunity.

Sree Balakrishnan:

You're very welcome, Neil

Neil Morrison:

Sree Balakrishnan is a partner and Zohaib Akhtar is a senior manager at Deloitte Canada. On our next episode, we speak with Cathy Cobey. Cathy is the EY global responsible AI co-lead, and she talks about the need to balance the potential opportunities of AI with the very real risks it presents. I

Cathy Cobey:

I think CPAs need to really start to appreciate how they can validate the reliance that they should place on the outcomes coming out of AI and how they can build some of that uncertainty into our processes where right now there is a very high expectation that when we provide information, it's got very high reliance and relevance and reliability, and yet AI doesn't always meet that threshold. But we don't want to then turn around and not use it. It's going to be a very valuable technology, so how do we balance those two sometimes competing priorities?

Neil Morrison:

That's Cathy Cobey, EY global responsible AI co-lead speaking on our next episode. That is it for this episode of Foresight: The CPA Podcast. If you like what you heard, please give us a 5-star rating or review wherever you get your podcasts and share it with your networks. Foresight is produced for CPA Canada by PodCraft Productions, and please note the views expressed by our guests are theirs alone and do not necessarily reflect the views of CPA Canada. Thanks so much for listening, I'm Neil Morrison.

Share Episode

Shownotes

Transcripts

Follow

Chapters

Video

More from YouTube