Key findings from CPA Australia’s 2024 Business Tech Survey

Garreth Hanley:
This is With Interest. A business, finance, and accounting news podcast brought to you by CPA Australia.

Gavan Ord:
Hello. Welcome to With Interest. I'm Gavan Ord from CPA Australia. In this podcast, we'll be talking about the artificial intelligence and business transformation insights gleaned from CPA Australia's 2024 Technology Report with Stephannie Jonovska.

Stephanie is an FCPA and is head of finance operations and transformation at BlueScope Steel, where she has extensive experience in implementing new technologies, driving transformation projects and human upskilling. Stephanie is also chair of CPA Australia's Digital Transformation Centre of Excellence. Welcome to With Interest, Stephanie.

Stephannie Jonovska FCPA:
Thanks, Gavan. It's great to be here talking about one of our favourite topics, technology applied in real life. And I'd like to extend thanks to all the members that completed the business technology survey. It's your input that helps CPA Australia deliver even more tangible value to members through the reports, webinars, articles, and podcasts like this. So let's have a great conversation, Gavan.

Gavan Ord:
Well, thanks, Steph. So the first question I have is around AI adoption and its benefits. The survey data shows that the main benefits from AI adoption are improved efficiency, improved productivity, and helping make better decisions. Could you discuss how you seeing AI benefit organisations?

Stephannie Jonovska FCPA:
I'll speak to my lived finance experience. AI-powered solutions are absolutely automating our routine financial processes such as accounts payable and receivable, reducing manual errors, and auditing our work, which is just fantastic. It's extra arms and legs, freeing up our people to do the work that humans do best, thinking, innovating, and creating. But what I think is more impactful is using AI to augment human decision-making in strategic areas through better data analytics capabilities, boosting our brilliance in effect.

And I'll provide an example. We've implemented machine learning, which is a subset of AI, with great success in our data collection forecasting process. It's now largely automated reducing the time taken to generate this from three days to just one hour with the entire thing happening in the background autonomously. But more importantly, our forecast accuracy has improved by 30% without people able to explain any variances at a customer level. Just imagine trying to do that Excel across thousands of customers.

You just can't do it. If there is anything extra people know about the forecast like weather or customer behaviour changes or logistic challenges, they overlay their human knowledge on the machine learning results. So their work is augmented by an at-scale machine learning model. They have more detail than ever to explain the results boosting their brilliance, and they still own the numbers. And I think this is just such a good example of the benefits of AI.

And our people are learning that human plus machine equals better, not one versus the other. So yeah, it's definitely improving efficiency. It's increasing productivity and boosting our brilliance. I think they're great tangible benefits of AI.

Gavan Ord:
I love how you said, and we've discussed this before, machine plus human equals better. It's not cut and paste from what the machine tells you.

Stephannie Jonovska FCPA:
You've got to overlay your human brilliance over it. Yeah.

Gavan Ord:
That's right. And you mentioned about the project you just did about debtors. So that's something you brought in external, or was it done by internals? How did BlueScope do that project?

Stephannie Jonovska FCPA:
Well, our approach is if something creates a competitive advantage to build capability internally, so partnering with IT, with the business, the credit teams, our finance teams, and my own team to craft a solution because that diverse brains brings a better outcome to our organisation.

So we built the capability internally, data engineering, data science, but importantly, making sure that the outcomes were accessible to our finance teams through things like data visualisation, Power BI. I think it's really important that our people are building capability so that they still remain relevant and marketable.

But also, seeing that this is not a threat, this is something that is helping us do our human work even better.

Gavan Ord:
Is this opening up opportunities for the staff that we're doing that to actually do more value-adding more interesting work?

Stephannie Jonovska FCPA:
Yeah, they can do more valuable and interesting work. They can upskill themselves. But they can also add more value to the outputs from the machine learning model. They don't have to spend the three days crunching it.

They can look at the outputs and then analyse it and give better insight and influence business decisions. So I think that's where the power is coming now. They're not doing the crunching, they're doing the thinking to overlay on that. And I think that will always deliver a better outcome for both our humans and our organisation more broadly.

Gavan Ord:
And of course, AI is not without risks. What are some of the obvious risks when using AI, bad data, corrupted data, bias, leaking sensitive information? Could you provide some tips on how businesses can manage AI risks and maybe mitigate those risks?

Stephannie Jonovska FCPA:
Yeah, I can. And I've always said, "Be partly excited and be partly scared." Makes you both curious and also cautious in this area. So you mentioned bad data, it leads to poor decision making. And to mitigate this, I strongly believe in investing in data governance frameworks to help ensure the data used by AI is accurate and up to date. This is a huge challenge for most organisations. It sounds like a lot of work that no one's excited by, but it's really critical work.

And a data governance framework involves policies, roles, responsibilities, data cataloguing, quality, management, privacy, security tools, technology, performance management, there's a whole raft of things that go into this. But my firm view again is the reason you implement data governance frameworks is to extract value from data. If you have good data, you're going to make better decisions and extract more value. And remember, data is the foundation to all AI projects. And I think data governance might be a good topic for future podcasts for our members. And you mentioned IP leaks and data security and things like that. And look, huge, huge challenge.

As AI models, particularly those trained on lots of data may inadvertently expose sensitive information. It's a huge topic, so I'll only cover a couple of things. On the input side, data minimization. Just use the necessary data to train the model. That's it. Data masking, remove the identifying elements from data. And then data access controls. Control strictly who can access training data, particularly sensitive data. And now, on the output side, output filtering, that could prevent the release of sensitive data. And model auditing, making sure that you're testing for potential IP leaks. So quite a complex area, this piece, but I think a really critical area.

And then, AI bias, lots of different views on this, but my view is it can be mitigated by ensuring diverse data sets are used during training. And by involving diverse teams. We often forget about the diverse teams in the development process. In that machine learning data collection project I mentioned previously, we did involve a credit team, business finance team, IT, and my digital finance and innovation team, all bringing different lived experience to the project, a different lens. In addition to that, I strongly recommend regular monitoring of AI outputs for fairness and for transparency to address any unintended bias in the system, and also to refine the model. It's not set and forget.

So what is working really well for us, is the diverse team I spoke about, meets monthly to discuss the results prior to the submission of the forecast. And it's an opportunity to review the results, feeding new information required into the model, and it creates really high ownership for the numbers. And I learned a new term, I'm going to share it with members here. It keeps the human on the loop eyeballing what is going on, adjusting what is going on, and then owning it. And I think that's just so critical with the risks that we're exposed with implementing new technologies like AI.

Gavan Ord:
I know we had this discussion at the Digital Transformation Centre of Excellence is that you still need to maintain that professional scepticism. You still need to run your eye by it and say, "Does this seem right?"

Stephannie Jonovska FCPA:
Yeah, yeah. As far as professionals, we still own the numbers no matter how it's generated. And success is human plus machine not saying, "Well, the machine told me that, and therefore we just accepted it." That is really a failure. And again, can't stress enough that our people must own the numbers. And I think that human augmentation is really, really powerful because that does give us access to better, wider, deeper information, again, to make very impactful decisions. So pretty exciting time, right?

Gavan Ord:
Yeah. But obviously, AI is not the only technology that's out there. What other emerging technologies do you believe will have a big impact on business in the next 5 to 10 years? We asked this question because there's a lot of hype around emerging technologies, so it's more around which ones do you think will actually have a real impact rather than those that are hyped to have a big impact.

Stephannie Jonovska FCPA:
This is a huge area and a huge topic. I will focus on AI though. I think AI by itself has a capital value, but I do think there is some transformational technologies that will become mainstream sooner than 5 to 10 years. And one of them is composite AI. Think of this like mixing different kinds of AI together to solve problems better than any single AI could do on its own. So more versatile. More resilient.

An example could be combining natural language processing or NLP for understanding customer questions, with machine learning to offer personalised answers, and automation to perform tasks like booking appointments for smarter chatbots. That's what people want. They want a very personalised experience. I think something like composite AI, even though in this AI, all my answers are about AI actually, Gavan. I think it's a really powerful one around the humanization and that personalization.

The other one is quantum AI, so combining quantum computing with AI, faster, more powerful algorithms. I think it'll revolutionise industries like finance for portfolio optimization, pharmaceuticals for drug discovery, and I already know that organisations like IBM and Google are working on quantum systems that could optimise advanced AI analytics faster than current supercomputers.

So watch out world, that's going to be massive. And my favourite, favourite that I think will have impact is causal AI, identifying cause and effect relationships. Not just correlations, which I believe will improve decision-making. So predicting how actions influence outcomes. It's really useful where AI needs to act autonomously, maybe a scary word for all of us, or adapt to changing conditions, but I think it could still be very powerful.

Financial institutions can use causal AI to optimise customer retention strategies or marketing campaigns by understanding the cause and impact of different actions on customer behaviour. We can see that human plus machine again coming out in all these examples. Although I've been all AI talk, the iterations of AI that I think will have an impact.

Gavan Ord:
And I mentioned hype. Do you have any tips on how a finance leader can manage the hype cycle? I think you described it as the shiny-itis, how do you manage that?

Stephannie Jonovska FCPA:
Shiny-itis, I think every day I get about five or six emails that say, "Can you buy this payroll system," or, "Here's an accounts payable solution that'll fix everything, every problem you've ever had." I manage it by staying very focused on the problem to be solved. What am I trying to solve? Then I will do good research on what are tried and true mechanisms or can we optimise our existing systems and sweat the asset if you like to solve those business problems.

And sometimes you don't even have to go to that. You might even just need to eliminate or standardise the processes to solve the problem that you're trying to solve. So I think staying focused on your problem and then seeking a solution, maybe it's digital where we use digital sensing or perhaps there's some process improvements to be made.

I think this is a real risk that there is so much hype that everyone just wants to implement something that involves the letters AI in it that we can get a little tripped. And I think that's really wasteful in terms of not only dollars but also our valuable time.

Gavan Ord:
As a senior finance leader yourself, what tips do you have for businesses on how best to manage their digital transformation journey?

Stephannie Jonovska FCPA:
Well, the first hot tip is to start with a solid digital strategy aligned with both short-term and long-term goals, and making sure it's business-led and digital-enabled, not the other way around. Like I said, often you see a strategy that says, "Implement AI." But for what purpose? What are you trying to do? Retain customers, grow market share, improve the employee experience, what are you trying to achieve by implementing AI?

And if you look at ourselves in our commercial team, our 2030 vision is to be an integral, innovative, and passionate team enabled by digital or to ensure we have the capacity and capability to be the awesome business partners we aspire to be. And I think that having that bigger vision is really important to start with. Secondly, it's crucial to prioritise talent. So AI by itself is pretty useless, but human plus machine is pretty amazing. So invest in upskilling or re-skilling your team, ensuring they understand how digital can augment their work, and again, boost their brilliance.

Many businesses in the survey reported a shortage of tech talent addressing this internally through training is generally more cost-effective than hiring externally and it can help our people stay very relevant and marketable and also increase the employee satisfaction. For us internally, our digital upskilling programme has kicked off, again, to ensure we can deliver on our ambition of being awesome finance business partners, but also to reduce the fear of failure, to reduce the worry about the impact on roles, and also the scepticism of the benefits.

And obviously, focus on cybersecurity. As businesses digitise, they become more vulnerable to cyber threats. Make sure it's just business as usual. Cyber incidents, if it will happen, it's going to happen, so be ready for that and put in the best controls that you can afford. And finally, one that I think is really hard but really critical is to measure progress continuously. Really hard to measure digital transformation success. You have to set clear metrics. Whether it's operational efficiency, customer satisfaction, and revenue growth, set them, then monitor them regularly.

And this allows you to make some adjustments to your efforts and make sure you stay on track. With our ambition of creating capacity and capability to be that integral, innovative, and passionate finance business team, we measure capacity creation. And we currently have a pipeline of about 24,000 hours of capacity creation for our people. That's just amazing, because again, repurposing that time to work with the business on strategy, to work on your own upskilling, to work on other projects that will create even more capacity or capability, I think is amazing. And the good thing is we're capturing all this via app, and then it's all monitored in a visualisation report.

And I think that's really living our digital-by-default ambition. So it's not in an Excel file filed away somewhere on a server, it's app-driven and visualised out, and gets reported to our CFO and our commercial leaders. In addition to that, we also report a happiness index, so we have a measure of employee experience. That's the differential between how the process was prior, and post the digital initiative.

And again, I think that makes it very real for our people. It's not a technology journey, it's a human leadership journey. So I feel very strongly about that but digital transformation, not a technology journey. It's a human leadership journey, and I feel personally very privileged to be a finance leader in this AI generation. It's a very exciting time for all of us if we embrace it.

Gavan Ord:
I agree. Thanks for information today, Steph. That was a really great conversation.

Stephannie Jonovska FCPA:
Thanks, Gavan, for the opportunity to share some of my lived experience with our members, and I hope that they found that valuable and can implement that in their day-to-day lives.

Gavan Ord:
I'm sure they did. For more information about the topics we've discussed in this episode and a link to CPA Australia's Business Technology Report, you can refer to the show notes for this episode. With Interest is a regular podcast. If you like today's show, you can subscribe on your favourite podcast app by searching for CPA Australia's With Interest. I'm Gavan Ord. Until next time, thanks for listening.

Garreth Hanley:
You've been listening to With Interest, the CPA Australia podcast. If you've enjoyed this episode, help others discover With Interest by leaving us a review and sharing this episode with colleagues, clients, or anyone else interested in the latest finance, business, and accounting news. To find out more about our other podcasts and CPA Australia, check the show notes for this episode. And we hope you can join us again for another episode of With Interest.