The future of accounting in an AI-driven world

Amantha Imber:
AI is not magic and it doesn't replace expertise. An expert with AI still produces better output than someone that is not an expert working with AI.

Anders Sormann-Nilsson:
There's a new Moore's law for agentic AI. So while you're sitting sceptical on the sidelines, your competition that's adopting agentic AI, they'll 2x you in the next quarter, 4x you in the next two quarters, and in a year's time, they have 16x-ed you in terms of their capacity and productivity output.

Patrick Viljoen:
Hello, everyone. My name is Patrick Viljoen, sustainability lead at CPA Australia. Today, we are unpacking the big forces reshaping finance, accounting, and business across Asia Pacific, from artificial intelligence and automation, to cybersecurity and sustainability reporting.

Joining me today are two experts with very different perspectives, innovation specialist Amantha Imber, and futurist Anders Sormann-Nilsson.Together, we'll explore what these bold signals mean for professionals on the ground, which skills will matter most, how roles are changing, and how to stay competitive in a rapidly evolving environment. Amantha, Anders, welcome to today's podcast.

Amantha Imber:
Good to be here.

Anders Sormann-Nilsson:
Great to be here.

Patrick Viljoen:
I want to start by getting a relatively short answer from each of you to an incredibly big question. So finance and accounting are being reshaped simultaneously by AI, cybersecurity threats, sustainability reporting, and then also rapid digital transformation. And all of this set against the existing geopolitical tensions that we're all experiencing.

From where you sit, what are the biggest forces changing the profession right now and why do they matter for professionals' day-to-day's work?

Amantha Imber:
Well, look, my lens on this is around knowledge workers and gen AI capabilities. So it's a specific lens that I work in at Inventium AI. And we are seeing just the complete redesign of how work is done. But the problem is that I feel like I see two completely different groups of organisations. There are those that are really redesigning how work is done. They're relooking at different workflows and where can AI help. And then there's also a lot of organisations that have just started using ChatGPT or Copilot and they're putting in kind of crappy prompts into their emails and producing AI slop that overuses the word "delve" way too much. And I don't see a lot in between.

So it's changing how some organisations are doing things, but then there are others that I would say are real laggards and are not changing at all.

Patrick Viljoen:
Yeah, no, that's definitely true. And we use Copilot quite a lot. And believe me, I've done some really bad prompting myself in my day.

Anders, your views?

Anders Sormann-Nilsson:
Well, I think we're really just at the knee of the curve of exponentiality at the moment where things are about to get really, really hectic. And we're certainly seeing some fast adopters, some innovators, but of course other organisations are real laggards as well.

I think the big hope here and the big promise of artificial intelligence is that it's going to help take the robot out of the human, and certainly in your profession, take the robot out of the human so we can do less of the menial and the mundane and actually focus more on the meaningful and the humane, which is the work we all want to do, which is the value added work. And I think this has huge repercussions for any CPA as they're thinking about their career, be they an emerging leader or be they more senior in an organisation or even be their founder of a company. This has huge potentiality.

I mean, even as a futurist, I have to hack my own workflow every day. And maybe I'm a little bit old school like this, but I've got a big yellow post-it note on my desktop, which reminds me every day to think AI first. And there's a great website called, theresanaiforthat.com, which literally, if you're interested in hacking your workflows and becoming more productive, a great point to start to rethink your workflow's courtesy of artificial intelligence.

Patrick Viljoen:
Yeah, thanks for that. Being a proud Gen X, I've been known to have sticky notes on my iPad whilst I'm trying to sort of navigate all these things. So yeah, it's not always pretty.

Amantha, you're on record for talking about AI fluency as a competitive advantage. And I think that's something that our members are also wanting to get their heads wrapped around. So beyond simply just using tools like ChatGPT, or in CPA Australia's case, Copilot, what does true AI fluency look like in practice within organisations?

Amantha Imber:
Well, look, with the organisations that we work with at Inventium AI, we see that there's two steps. Most organisations know that people using the tools and then knowing how to do the basics, like have a conversation with the AI where you're getting useful output, which is more than a lot of people are doing, we would call that AI literacy. And that is what a lot of organisations are focused on capability building around. But then they stop and they're like, "Great, we can tick that box. We are done. Let's move on to the next thing." But the thing is you are not done. We say you need to first focus on literacy.

And then focus on leverage because most people do not think like a workflow architect or a process architect. They're like, "Great, I can write better emails. I can analyse this data in Excel better. I can write this report more quickly." But ultimately where AI is going to have the biggest gains for all of us is where we look at the workflows that we are doing daily, weekly, monthly, and we really think about where should AI and could AI play a role? And then we really advance our skill level and go, "Great, how do we redesign that workflow?" So I would say for the more advanced organisations or certainly professional services firms that we work with, first, we're doing literacy, but importantly, we're then doing leverage to get those really big productivity gains.

Patrick Viljoen:
Yeah. And I think it's a good point you raised on about rethinking the way that you actually do work and how that can actually unlock value. And I think that's something which for finance teams is going to be crucial as they engage with us into the future.

Anders, looking ahead three to five years, which emerging technologies or trends do you believe will perhaps fundamentally reshape how finance teams operate?

Anders Sormann-Nilsson:
Well, I think we've all been obsessing about all the potentiality of generative AI for a number of years now since ChatGPT came onto the market. But I think the big shift that we've seen in the last 12 months and even now continuing into the future is the rise of agentic AI. And there's a new Moore's law for agentic AI. So this is not science fiction, this is just pure maths. The new Moore's law for Agentic AI is that rather than doubling in its capacity every 18 to 24 months, which is the old computing laws, it's now doubling in its sophistication every three to seven months. So if we take the more extreme scenario here, while you're sitting sceptical on the sidelines, your competition that's adopting agentic AI, which also executes for you and executes for CPAs, they'll 2X you in the next quarter, 4X you in the next two quarters. And in a year's time, they've 16Xed you in terms of their capacity and productivity output.

So Agentic AI is a massive piece of news and I think it's something that we'll see more of. I was just speaking at a large payments conference yesterday in Sydney and I talked about the fact that at Banco Santander at the moment, together with MasterCard, they're now enabling agent to agent commerce and payments. Now, I don't know about you guys, but I think everyone gets a little bit nervous when they give their kids their first credit card, or of course you're the signatory on that, or even giving your loved ones another credit card in your name that you're backing. So imagine the potential, not just for greater commerce and faster commerce, but also the opportunities for bad actors to cyber hack the fact that now also your AI agents are going to have credit cards and interacting with one another.

So I think that's a big piece of news. We can certainly touch on other trends, but I think agentic AI is one of the things that we'll all have to get used to because of course the workforce is not just going to be humans. We'll also count our workforce and our headcount in how many AI agents are we now employing as well.

Patrick Viljoen:
Yeah. And that's going to make for a very interesting comparison of FTE over duration and time. Amantha, I just wanted to come back to a point you made, and we spoke about re-imagining work streams, but do you see AI primarily augmenting professionals' judgement and productivity or replacing significant parts of their work, and particularly focusing here on routine accounting tasks? So we've spoken about the workstreams, but how can it actually influence how our members need to bring that level of judgement to their roles that they operate in?

Amantha Imber:
Yeah. I mean, human judgement is absolutely critical. And even just like in some research that I was reading a couple of days ago, AI allows those that are kind of below average or have less experience in a field to produce something of much higher quality. But we know that there is actually a wall or like a ceiling, I guess, on that skill level in that an expert with AI still produces better output by quite a significant margin than someone that is not an expert in that particular domain working with AI.

So AI is not magic, it is very skilled, but it is absolutely not magic and it doesn't replace expertise. Where things will be in a year from now, I have no idea. And certainly the models that most of us would be using, whether that's ChatGPT or Claude or Gemini, Copilot, whatever, I mean, every month there are significant improvements. And if, for example, particularly Copilot, a lot of our clients use Copilot, if you tried using Copilot six months ago and then you did nothing and then you used it today, it's a vastly different experience, which just speaks to how quickly these LLMs are improving at what they do and the output that they're giving you.

Patrick Viljoen:
Well, it's really valuable to know. Anders, you made a point about having agents on both sides of a value chain speaking to each other. Now that in and of itself, for me, raises some concerns, but that's just me. But I think looking back in terms of that kind of connectivity between organisations where AI is actually fulfilling that gap or bridging that gap in between two organisations, in my mind, it also raises the spectre of cyber threats that can potentially impact on that. So if we cast back on technology or this particular piece of technology also expanding into the cyber threat landscape, how should our leaders actually balance innovation with risk management and securing or safeguarding resilience of the organisations?

Anders Sormann-Nilsson:
A couple of points on that. So I've just come back from Las Vegas where I keynoted for 1,200 internal auditors, ironically in Sin City, I should say, on responsible AI and the ethical deployment of artificial intelligence. I think your question is spot on.

And I say three things to that. One of my main key points in Las Vegas was the fact that when we deploy artificial intelligence, firstly, we need to be able to explain the AI. In other words, we have to have explainable or X AI. Secondly, we need to be able to defend it should our organisation and up on the front pages of the Australian or the Sydney Morning Herald or the Australian Financial Review, like some accounting firms might have done recently because they produced AI slop or hallucinating reports for the government, for example.

And then thirdly, we need to be able to sustain our AI investments as well. I think the major thing here today is that for good and for bad, it used to be the case that you had to be a computer scientist or to be able to code, to interface with artificial intelligence, and that the coolest coding languages are no longer Java or Python, right? Today, the coolest coding language is Aussie English, or maybe Swedish Australian English, right? We've all become coders because AI now speaks human. And again, this is a problem and it's also a great opportunity. It removes the barrier for anyone to start deploying this. And it really comes down to a question of human creativity and creativity in the CPA profession.

But beyond this, just to give you a cybersecurity aspect to this, is that when AI speaks human, we can all create new cyber threats. And one example of this is that the engineering firm Arup, which in Hong Kong a couple of years ago, an employee will send a request supposedly from the CFO to authorise transactions of 25 million USD into 11 different Hong Kong bank accounts. Now, the employee in that particular instance did the right thing from an internal audit perspective. He set up a Zoom call with the CFO and a number of other people to authorise and see whether this was actually an accurate request as opposed to a deepfake.

Now, the deepfake courtesy of the artificial intelligence was so good that on that Zoom call, the employee bought that the CFO was on the call along with a number of other colleagues and executed the $25 million worth of transactions, which of course has been a majorly published piece of a deepfake scandal. So that is a cybersecurity threat that we're now living with and have to deal with in any organisation.

Patrick Viljoen:
And again, I just wanted to extrapolate that a little bit further, and this might sound like a bit of a curveball question. When our members tend to engage with AI, and from my background working within sustainability, AI is increasingly being used for extrapolation and forecasting of specific things.

So if we think about climate scenario analysis, for example, or if you're thinking about budgeting and forecasting into the future, one of the things we keep on hearing, particularly also from the assurance community is to say, "Well, AI for them still appears to be this black box paradox of producing an answer, but if they want to interrogate that or the assumptions that went into that particular answer, they still struggle to understand where they should be focusing and what they should be looking at."

So for the both of you, how would you address that kind of, I think, hesitance and trepidation about what is being bandied about in the market as the black box paradox?

Amantha Imber:
Well, I mean, the obvious place to start is ask your AI what assumptions it made. I mean, and often, depending on what tool you're using, you can actually see it's working. So generally it's a little dropdown menu to look at how it got to the conclusion, but I mean, ask it and it will tell you in plain English.

I am a really big fan of just making time to experiment with AI daily. The rate of improvement and the new features that drop all the time mean that the advice that we're giving today around, "How do you use gen AI? How do you get the most out of it?" could be totally irrelevant by the time that this podcast goes out. Hopefully it won't. Hopefully we're still good and we're still adding value. But in all seriousness, I mean, those that are taking even just 15 minutes per day to experiment and play and learn, you're going to be ahead of 99% of people because most people are not doing that.

Patrick Viljoen:
Yeah. Andres, any thoughts on that?

Anders Sormann-Nilsson:
Yeah, I think there is a massive first mover advantage in this space. And as I said before, it really comes down to human curiosity and human creativity. I call this the AI growth mindset with a quick cap nod to Carol Wick's work at the University of Stanford School of Psychology.

She wrote a great book called The Growth Mindset, which basically argues that people with a growth mindset as opposed to a fixed mindset are those that thrive into the future. Fixed mindset is, "This is my best work and this is ... I’ve peaked in my career." Capabilities, for example, I think now pivoting to an AI growth mindset just asks us, have we just scratched the surface of our human potentiality now that we're all becoming cyborgs essentially?

So I think in terms of the future of work and how we'll do work, we'll all need to adopt this AI growth mindset. We're also just seeing the huge potential here with mandatory sustainability reporting, which you kind of touched upon here as well a moment ago, another huge opportunity for us to think about not just the impact of artificial intelligence, but also just with the environmental costs of AI as well.

Now, if we think about the power hungriness of AI, we're going to be 4Xing the amount of power by 2030 that these AI data centres will be requiring. But also the amount of water that these data centres that are being proposed in Australia, for example, they could consume up to 25% of Sydney's water to just cool the data centres. So of course, with all these solutions that we can put to AI to use and some of the wicked problems that potentially could help us solve, it could also burn the planet at the same time.

So we have to be cautious from a sustainability reporting perspective that yes, we take a sort of twin transformation approach to this to invest in digital transformation and sustainable transformation at the same time and actually power these data centres with renewables as well. So it's a challenging but fascinating project for any smart CPA that's looking to position themselves well into the future.

Patrick Viljoen:
Perfect. Thank you.

Amantha, just coming back to you, we spoke about the drudgery of work being taken away. So back in the day in terms of journal posting and all that kind of stuff, which can be automated and work streams being reimagined. But on the back of that, we're also hearing that hiring patterns are effectively changing, so especially for entry level roles. And I wanted to ask you how automation is reshaping career pathways into the profession and how members need to negotiate that kind of problem as well.

Amantha Imber:
Well, I think that this is a really big problem because I don't think that many organisations are thinking, "What is my succession plan?" They're thinking, "Well, I can cut some costs by halving the number of grads that I can take on this year because the AI can largely do the job of a lot of the things that we get grads to do." But then what's going to happen in 10 years from now where you've got no one that you've put into that succession plan to mentor and nurture? So I actually think this is a leadership question, not a, "What should I do if I'm at the entry level?" And if you are at the entry level, look for organisations that are actually thinking about the future of their business and making sure that they do have a really strong talent pipeline.

Patrick Viljoen:
Yeah, 100%. Maybe just by way of a final question, if our listeners wanted to future-proof their careers, we're all about future-proofing the profession and making sure that our members can thrive into the future, what is one practical action or mindset shift that will make that big difference?

And Andres, I know you alluded to these kind of mindsets that need to shift previously, but what are the practical kind of things that our members can actually do?

Anders Sormann-Nilsson:
Yeah. So I'll give you a little bit of Anders' futurist boy maths here. So a very simple formula for success for the future, I believe, is to focus on three different types of cues or intelligence, right? And they're very humanistic, human-centred types of intelligences as opposed to artificial intelligence. And the first one is to really focus on our emotional intelligence, our EQ. It's something that we can nurture through neuroplasticity, through our entire lives, whether that's downloading a mindfulness app to help regulate our nervous system when we're dealing with staff or challenging situations.

Secondly, focusing in on our unique or our UQ, our unique intelligence as a unique psychographical one. In other words, really identify what's your superhero skill, but also be mindfully aware of where your weaknesses are as well, because some of those weaknesses are things that you actually want an AI to help augment you in.

And finally, to focus in and nurture our curiosity and our CQ, our creative quotient, right? And I'll give you an example of this in just a moment, but on top of those three things that we can really focus in on constantly learning throughout our lifetimes, sprinkle some digital augmentation and artificial intelligence as well, and you'll make sure that you have the most empathetic and also the most productive and augmented personal leadership brand, but also hopefully for your firm, the most empathetic brand moving into the future.

And I'll give you one quick example of an entrepreneurial mates in the entrepreneur's organisation where I'm a member who did this magically, right? He knows very well that his unique intelligence is not in negotiations. In fact, he says culturally he has problems asking for money. So what did he do? Rather than trying to become a really good negotiator, he downloaded Chris Voss's book, Never Split the Difference. Chris Voss is an FBI hostage negotiator.

So essentially, my colleague has created a negotiation AI agent in Claude so that whenever he's negotiating contracts now, he's using his CQ, that's his creativity quotient, he's using the emotional intelligence from Chris Voss, and realising that he has some weaknesses in negotiations, but he can augment his emotional intelligence in his communications with his clients to actually strengthen his margins. And this comes back to him being a fast adopter and digitally augmenting himself because now he’s leaps and bounds ahead of some of his clients and being able to actually negotiate better deals for him and his company.

So I think that's a great example of actually doing EQ, UQ, CQ together with digital augmentation. I would encourage all CPAs to do something similar.

Patrick Viljoen:
Wow, that's amazing. Amantha?

Amantha Imber:
Yeah, look, I mean, I absolutely think all those skills are important, but I think if your organisation is not investing in upskilling you in how to get more value from AI, then take matters into your own hands. I am obviously very biassed because I run an AI training company at Inventium, and we do great work, but just get it from somewhere. What we know from research is that you need a good five hours' worth of training to see a difference. Ideally, over the course of the year, about 81 hours' worth of formal training is what you need to really, really level up.

And for myself, I think about my own capability now as a very heavy user of AI compared to the me that I was in mid 2022. And there's no way that I could be running two businesses, hosting two podcasts, having written my fifth book, and keynote speaking all over the place without the help of AI. And I am very much a fan, obviously, of the human doing a lot of the thinking, but AI is what allows me to get massive gains and a massive return on my own time.

Patrick Viljoen:
Amazing. Well, plenty to reflect on there. And it's clear I think that the profession is entering a period of profound change, but it's also a significant opportunity for those willing to adapt. And I think that came across quite strongly from this conversation as well, from building AI fluency and strengthening cybersecurity awareness, to embracing sustainability as a driver of value. The message from our guests is that the future will reward curiosity, judgement , and more importantly, continuous learning. My thanks to Amantha Imber and Anders Sormann-Nilssonfor sharing their insights with us today.

Anders Sormann-Nilsson:
Thank you, and see you in the future.

Patrick Viljoen:
For listeners who would like to explore these trends further, you can find links to the related resources in the show notes. You can also check out INTHEBLACK's 140th anniversary special edition online, which has a feature length article on the topics we discussed today. Also, do not forget to subscribe to INTHEBLACK and share this episode with colleagues across your network. Until next time, thanks for listening.