How can businesses stay ahead of the rapidly evolving risks of AI-driven fraud? In this episode of Tech Talks Daily, I sit down with Bradley Niedzielski, Audit & Assurance Partner, Digital Transformation at Deloitte, to explore the growing threat of generative AI in financial fraud and how organizations can strengthen their defenses.
With deepfake technology, synthetic identities, and AI-generated forgeries becoming more sophisticated, traditional fraud detection methods are struggling to keep up. Deloitte projects that AI-driven fraud losses in the U.S. could reach $40 billion by 2027—more than tripling from 2023.
Bradley shares insights from Deloitte's latest research, including a startling statistic: over 25% of businesses have already faced a deepfake financial fraud attack, and more than half expect these threats to escalate within the next year.
Despite this growing concern, only 7.4% of companies prioritize technology as their primary defense, and less than half feel confident in their ability to combat AI-powered fraud. So, what's holding them back? And what steps should organizations be taking right now to strengthen their fraud risk management frameworks?
We discuss the critical role of AI governance, the importance of transparency in AI decision-making, and why human oversight is still essential—even as AI becomes more autonomous. Bradley also highlights how finance departments increasingly embed AI into their workflows, from automating financial close processes to leveraging AI-driven scenario planning. But as AI becomes a core part of financial operations, ensuring data integrity and preventing blind trust in AI-generated outputs is more critical than ever.
Beyond risk mitigation, we explore how businesses can prepare their teams for the AI-driven future. Bradley shares best practices for upskilling finance professionals, fostering a culture of innovation, and ensuring AI adoption aligns with broader organizational goals. With AI reshaping finance functions at an unprecedented pace, how can companies implement it securely while maintaining regulatory compliance and operational efficiency?
As generative AI continues to disrupt the financial landscape, how can businesses balance innovation and risk management? And what strategies will define the next generation of AI-powered finance operations? Tune in to hear Bradley Niedzielski's expert take on navigating the opportunities and challenges of AI in finance.
[00:00:04] How is AI reshaping the finance department of tomorrow? Well today my guest is going to join me in exploring how generative and agentic AI are beginning to transform the way finance teams are operating. Offering both significant opportunities and maybe a few complex challenges along the way. From automating workflows to tackling fraud risks,
[00:00:28] my guest at Insights will offer a roadmap for navigating the integration of AI into financial processes. We'll discuss the shift of traditional automation and autonomous AI agents, the importance of governance and human oversight and why data quality is the foundation of successful AI adoption. And maybe even shed light on how finance professionals can prepare for this transformation.
[00:00:55] Who is my guest? Where is he from? What does the future hold for AI and finance? And how can organizations ensure that innovation goes hand in hand with security and transparency? Enough teasing and scene setting for me. Let's get my guest onto the podcast now. So thank you so much for joining me on the show today, Bradley. For everyone listening, could you tell them a little about who you are and what you do? Yeah. Well, first off, thank you, Neil, for having me today. Really look forward to our discussion.
[00:01:25] And, um, you know, really what I do is on a daily basis, I work with our finance departments of various organizations and I assist them with, um, getting the most out of their people, the processes and technology that they have. Well, obviously over the past couple of years, we have really seen a focus on technology, especially within the generative AI space. And I'm curious as someone that spent so much time in this space, how have you seen the rise of
[00:01:52] generative AI over the last, what, two, three years? Have you seen it change the landscape of things like fraud, particularly when we look at the emergence of synthetic identities, deep fake impersonations and document forgery? These are just a few of the things off the top of my head, but what are you seeing here? Yeah, the rising generative AI has introduced powerful tools for fraudsters, including the ability to create realistic deep fakes, forge documents and impersonate
[00:02:22] executives in high stakes scenarios. So finance and accounting departments, they face elevator risks as deep fake enabled fraud targets. Uh, they, they target processes like identity, identity verification, financial approvals and transaction validations. Um, we've seen that organizations can mitigate these risks by enhancing security protocols, such as implementing multi-factor authentication,
[00:02:49] increasing layers of approval and leveraging gen generative AI powered fraud detection tools. We're essentially going to be fighting AI risk or AI fraud with AI as well. Think of it also where we've implemented virus technology in the past, or else you're going to implement generative AI technology to help identify some of that fraud. Now proactive workforce training is also essential
[00:03:14] to help employees recognize and respond to AI generated fraud schemes, along with interdisciplinary collaboration across departments like IT and HR. Sharing lessons learned from thwarted attacks can help protect the broader business community and strengthen collective defenses. So a proactive risk management framework that integrates generative AI detection tools and regular regularly evolves in
[00:03:41] response to emerging threats will be crucial to reducing vulnerability and finance. Another question I'd love to ask you, Bradley is we're at the beginning of a new year and people are looking at managing things differently, doing things differently this year and trying to determine some of those new trends on the horizon. So from your perspective, what's your outlook for AI innovation and AI innovation that you think will
[00:04:05] most impact finance departments this year? Yeah. And if we look back, the last few years have been defined by testing and adoption of gen AI tools and really just a proofs of concept of what gen AI can possibly do. But in 2025, AI will be used in your finance department on some scale as it has become embedded in many platforms and systems that organizations use today. The large language models or the LLMs behind gen AI
[00:04:35] are more capable of understanding context, generating human-like text and performing complex tasks. In 2025, we may see this evolution from traditional automation and LLM based applications to autonomous AI agents, also known as agentic AI. This agentic AI can reason, plan, remember and act autonomously. While unlike typical language models that
[00:05:03] automate individual tasks, these AI agents can automate entire workflows and processes. This represents a shift in AI being deployed as a supported tool to being leveraged as one that is more collaborative. So Deloitte predicts that in 2025, 25% of companies that use gen AI will launch agentic AI pilots or proofs of concept. And that number will grow to 50% in
[00:05:31] 2020. So agentic AI could become a cornerstone of finance transformation, unlocking faster, more efficient finance processes and empowering more strategic forward-working decision making. Governance and human oversight will remain integral to finance and accounting workflows, ensuring processes maintain the high standards of accuracy
[00:05:54] and transparencies that these functions demand. When it comes to implementing AI, companies have the option to build, buy or utilize existing systems. They can build their own AI platform or they can buy a license for AI-powered software. But many organizations don't realize they can use AI that is already integrated into their existing
[00:06:18] tech solutions. For instance, in a finance department, this could involve making use of AI capabilities already built into the reporting software. And we are approaching three years with mainstream adoption in AI. So with that becomes a fair amount of maturity. And we are seeing the rapid deployment of AI within finance functions now. So how do you see governance changing? And how does it stay the same as well? And why are these factors so important?
[00:06:46] Yes, governance frameworks, they must adapt to ensure that AI outputs are traceable, auditable and meet regulatory expectations. Transparency in AI decision-making will remain non-negotiable. While innovations like agentic AI increases efficiency, finance departments must still implement robust preventative and detective
[00:07:09] controls to validate data integrity and ensure process accuracy. The need for a human oversight will also evolve over time, with the responsibility shifting to individuals monitoring AI systems rather than actually performing the manual tasks. So organizations should ask themselves, does the control have an AI
[00:07:32] element? And how does the human in the loop get comfortable with the output and input of the AI? Blind trust in AI is the biggest risk out there. Companies will need to ensure their AI solutions meet compliance and accountability standards while integrating tools that track exceptions and document actions for audit readiness.
[00:07:53] And a few moments ago you mentioned agentic AI and it does seem to be one of the new applications of AI and it is widely tipped to be the big buzzword of 2025. So there's gonna be a lot of business leaders hearing about agentic AI. But one of the things I try and do on this podcast is try and work out more or try and explore the real value that are ROI and what does it mean for businesses.
[00:08:18] So for anyone listening doesn't heard all the hype around agentic AI, what are the actual benefits that financial departments could derive from these new capabilities? Yes. At agentic AI it enables the automation of the end-to-end workflows while ensuring tasks like account reconciliation, journal entry creation, and reporting are completed faster and with fewer errors. So by
[00:08:42] accelerating processes like the financial close, valuable information can be delivered to the users of the financial statements on a more timely basis. Finance teams can then focus on strategic priorities such as scenario planning and forecasting rather than just simply a retrospective reporting. AI-driven task management improves operational efficiency, allowing for real-time exception handling, and better resource allocation. So the
[00:09:12] integration of generative AI enhances productivity by enabling seamless collaboration across tasks and departments. So I guess the big question that a lot of people listening will be asking is how can their company prepare for and optimize for that implementation of agentic AI? What role does data play in this? That's something that often is something we don't talk about enough, I think. But what do you say here?
[00:09:41] Yes, I always stress that data quality is foundational. Core data inputs can magnify errors and AI outputs. So this underscores the importance of pre-deployment testing and robust implementation processes. So as with generative AI, the organizations that have the most robust and best managed data inputs will reap the most benefits from
[00:10:05] agentic AI. Overall, having a human in a loop is still important, even with agentic powered automations. Companies should adopt a hybrid approach. So combining automated processes with human oversight to ensure accuracy and regulatory compliance. Also, fostering a culture of innovation and equipping teams with the skills to utilize and
[00:10:30] develop AI effectively will be critical for the long-term success of organizations. Any new application of AI, if it's generative or agentic, should be accompanied by commensurate investment in secure cloud-host environments with encrypted data flows to mitigate security risks and enable trustworthy
[00:10:50] AI adoption. I think we will increasingly be talking about maturity of AI throughout 2025. So again, for people listening, how can finance departments ensure that they're staying ahead of that AI maturity curve? Do you think things like talent and culture are important to this? One of the reasons I say that is in my former
[00:11:14] IT life, I often saw new tech projects and things being thrown over the line without taking into account talent and culture. Is that more important than ever, would you say? It absolutely is. And finance departments should align their AI strategy with the business objectives. So ensuring that technology adoption serves the broader organization goals. To do this, they should develop an AI-ready
[00:11:42] culture by prioritizing continuous learning and upskilling employees to adapt to emerging technologies. Using sandbox environments for initial testing, this can minimize risk and ensure AI systems align with operational and compliance requirements. And also get existing employees comfortable with the technology. Companies should absolutely empower the new generation of finance professionals who are already familiar with
[00:12:10] AI technologies. They will drive more seamless adoption over time. A significant amount of the workforce today grew up with the internet ingrained in their everyday lives. The next generation of finance professionals are interested in entering the workforce. They will have grown up with AI ingrained in their everyday lives. And we should reach for that group of individuals and really use their experiences that they have.
[00:12:37] And at the very beginning of our conversation, we were talking about some of the fraud risks there. Anybody interested in implementing AI, whether it be agentic AI or anything, is there anything else that they should be aware of, be cautious about? Maybe it's the responsibility of doing that or the ethics or anything, anything else that you think they should be aware of?
[00:13:00] Yeah, and I'll jump into what we discussed a little bit earlier, but the rise of generative AI has introduced powerful tools for fraudsters, including the ability to create realistic deep fakes, forge documents and impersonate executives in high-stakes scenarios. And we've actually seen this in practice already. Finance and accounting departments, they face elevated risk as deep fake enabled fraud targets,
[00:13:25] processes like identity verification, financial approvals and transaction validation. So organizations can mitigate these risks by enhancing security protocols, such as implementing multi-factor authentication, increasing layers of approval and leveraging Gen AI-powered fraud detection tools. Proactive workforce training is essential to help employees recognize and respond to AI-generated fraud skills,
[00:13:53] along with interdisciplinary collaboration across departments like IT and HR. Sharing lessons learned is probably one of the most valuable sources of information we have. So share the lessons learned from thwarted attacks that can help protect the broader business community and strengthen collective defenses.
[00:14:14] Proactive risk management framework that integrates Gen AI detection tools regularly involves in response to emerging threats will be crucial to reducing vulnerability in finance functions. And I think many businesses will be looking towards you and indeed Deloitte for guidance as they're looking to implement AI. And there is a real, a very real pressure on everyone to be in a state of continuous learning right now.
[00:14:42] I think we all feel it, but I'm conscious. As someone right in the heart of this, someone that's leading the way, where or how do you self-educate? Any tips there you can share? Well, I'll tell you, I'll bring a more broader perspective to that. So we're in the information age right now, which is fantastic, but can be overwhelming because we have access to so much information at almost zero cost, but we simply don't have the time to learn and consume everything.
[00:15:12] So personally, I go back to basics. I talk with people every chance I get, and I have genuine, meaningful conversations. And I ask questions about their experiences within finance, if it's the new hire within the organization, or if it's the person who's been there for 30 years. These conversations generally lead to an aha moment for me, and I only either have a new topic or a tool to research and learn about,
[00:15:39] and or maybe I'll make a slight change in the way that I perform my own job or the way that I work. I would stress, though, that this can be from all works of life. So to this day, I learn from my parents' experiences, from my wife, from my children, and many friends and strangers who aren't even in the same field as me. Different perspectives and different experiences will help you learn as a professional.
[00:16:02] Now, people and information may be abundant, but one thing that you don't have is an unlimited amount of time. So many years ago, I asked a mentor of mine how he keeps up with all of his email traffic and other workplace administrative items, because as soon as we step into the office nowadays, it almost feels like we have back-to-back 30-minute meetings. So he told me he gets in earlier than everyone else, and he feels like he gets a head start on everyone else.
[00:16:30] And he specifically blocks time in his calendar for reading or for learning. And no matter what, unless it's an extreme emergency, he will not give that time up. So this is advice that I took, and I blocked certain time during my day for that exact task. But the other thing I did was I also blocked 30 minutes. For me, it's every Thursday to call people in my virtual Rolodex to check in and see how things are going. Ask what they're working on and what is top of mind for them.
[00:17:00] But some Thursdays, I just talk to a voicemail or I hang up and send it with text. But most Thursdays, I had a great conversation and I learned something new. Fantastic advice. I absolutely love that. And I'm conscious we've covered so much in a short amount of time today. So for anyone listening wanting to find out more information, dig a little bit deeper on some of the topics we explored together today. Where would you like to point everyone listening?
[00:17:25] Yeah, today I would point everyone back to Deloitte.com backslash US backslash audit AI. And that's a great starting point for you with materials of everything that we're doing here at Deloitte and everything that we're seeing happen in the future. And I'd welcome any conversation from there. Awesome.
[00:17:49] Well, I'll add a link to that to make it nice and easy for people to find out more information and dig a little bit deeper on some of the things we talked about today. Will it be the escalating risks of Gen AI enabled fraud, business vulnerability, the workforce gap in fraud preparedness and some of the good stuff and actionable steps about implementing agentic AI too. We could have talked about this for another few hours, I'm sure. But thank you for starting this conversation.
[00:18:17] I'm hoping people listening will find this really valuable. And I urge them to check out that link I'll put on the episode. But thanks for joining me today. Thank you, Neil. Appreciate it. Thank you to all the listeners. I think it's evident that the integration of AI in finance isn't just a technological shift. Having spoken with Bradley today, I think it's a fundamental evolution of how teams operate and create value.
[00:18:41] From the benefits of agentic AI to the critical need for governance and risk management, I think my guest provided a clear vision of how finance departments can embrace AI responsibly. And the road ahead will require continuous learning, collaboration across disciplines and a steadfast commitment to data quality and transparency. We shouldn't ignore these things. But what are your thoughts on the role of AI in finance?
[00:19:08] Are you prepared to balance the opportunities with the risks or do we need to rethink our approach to this transformation? I'd love to hear your thoughts wherever you are at right now. Please share it with me. Connect with me. LinkedIn, X, Instagram, at Neil C. Hughes. Don't just hit follow. Send me a quick note. I'd love to hear your perspectives. But that's it for today. So I'll be back again tomorrow with another guest. But remember, stay curious. Keep exploring the future of technology in your business.
[00:19:37] And we'll keep this conversation going. Bye for now.