Fraud Talk Podcast: Navigating the Intersection of Technology and Compliance: A Conversation with Christian Focacci, CEO of Threat.Digital

Intro

Rihanna: Hello and welcome to Fraud Talk, the ACFE's monthly podcast. I'm Rihanna Scoggins, the Content Manager for the ACFE, and today I'm joined by Christian Focacci, CEO and founder of Threat Digital, and an expert in compliance technology and data management. With a background in leveraging data and machine learning to tackle risk and compliance issues, Christian brings a unique perspective to our discussion. Thank you for joining us today, Christian.

Christian: Great to be here, Rihanna.

Background and Motivation

Rihanna: Before we dive into our deep conversation, could you share a bit about your background and what motivated you to focus on financial crime compliance and risk management?

Christian: Certainly. I started my career as an investigator in large banks, working on enhanced due diligence and anti-money laundering investigations. As technology became a crucial part of these processes, I transitioned to the tech side. My motivation stems from realizing the potential of technology in streamlining information identification and processing. This dual perspective allows me to bridge the gap between the investigative and technological aspects of the industry.

Founding Threat Digital

Rihanna: Now, as the CEO of Threat Digital, can you tell us about Threat Digital and what motivated you to found it?

Christian: After the success of my first company, which used machine learning to identify risk information from open-source data, I saw the transformative potential of new technologies like transformers and large language models. Although I initially wasn't planning to re-enter the compliance tech world, the capabilities offered by these advancements convinced me to start Threat Digital. We launched our product in June, leveraging machine learning and AI to address compliance challenges. The reception has been remarkable, with our technology providing a significant advantage in the market.

Leveraging AI and Machine Learning

Rihanna: Speaking of AI and machine learning, how is Threat Digital leveraging these technologies, and what trends do you observe in how other companies are using them?

Christian: The exciting aspect for us is the use of large language models, like those powering ChatGPT. These models excel in processing unstructured text, providing context and understanding. While many companies claim to use AI, the key is how they leverage these language models. The most impactful applications are when companies integrate language models from the ground up, making them an integral part of their product. This ensures meaningful outcomes and sets them apart from those simply adding AI as an additional feature.

AI Washing and Its Challenges

Rihanna: You mentioned AI washing earlier. Can you elaborate on what AI washing is and the challenges it poses?

Christian: AI washing refers to companies overstating their use of AI in marketing materials without delivering genuine innovation. It's a challenge when marketing departments prioritize AI as a buzzword, leading to inflated claims. This can mislead consumers, making it difficult to discern whether a product truly leverages AI or if it's just a marketing tactic. As someone deeply involved in AI, it's crucial to ensure transparency and authenticity in its application.

Challenges in Implementing AI in Compliance

Rihanna: Moving to the challenges organizations face when implementing AI in compliance, do you see resistance, and how do you address concerns about the black box nature of some AI systems?

Christian: Implementing AI in compliance faces a mix of interest and skepticism. While everyone is intrigued by AI, resistance arises from the inertia of changing established processes. Additionally, concerns about the black box approach, especially in investigations and fraud detection, are valid. It's essential to prioritize explainability in AI systems, ensuring that decisions are transparent and understandable. Overcoming these challenges involves a holistic approach, focusing on quality data, tuning the AI, and implementing post-processing measures for consistency and reliability.

Developing AI-Driven Solutions in Compliance

Rihanna: Delving into the development of AI-driven solutions, what challenges do you face, especially in compliance, and how do you build guardrails to prevent issues like hallucinations in AI outputs?

Christian: Developing AI-driven solutions in compliance requires mitigating the black box approach, ensuring explainability. We start with a holistic approach, focusing on data quality from the beginning. By integrating various techniques, including heuristic rules and other machine learning methods, we lay the foundation for high-quality data. Guardrails are then implemented throughout the process, from fine-tuning the AI to post-processing outputs. This meticulous pipeline ensures consistent, dependable results, preventing issues like hallucinations and maintaining the reliability of the system.

Advice for Companies and Technologists

Rihanna: Based on your experience, what advice would you give to companies or certified fraud examiners looking to implement AI into their compliance strategies?

Christian: Understand the basics of AI by exploring consumer tools like ChatGPT. Identify pain points in your processes and evaluate if AI can enhance efficiency. Start with a rough understanding of how AI works and then explore companies offering solutions for your specific use case. Engage in conversations, request trials, and ensure the technology aligns with your needs. Whether you're an early adopter or prefer a cautious approach, focus on solving genuine problems rather than adopting AI for its own sake.

Advice for Technologists Developing Anti-Fraud Solutions

Rihanna: Shifting to advice for technologists, especially those developing anti-fraud solutions, what guidance would you offer?

Christian: Solve your own problems. Focus on addressing challenges you personally encounter, as this authenticity resonates with others facing similar issues. Rather than adopting AI for the sake of it, develop solutions that genuinely enhance your processes. This approach ensures a deeper understanding of the problem, creating a more authentic and impactful conversation with potential users.

Closing Thoughts

Rihanna: Thank you, Christian, for sharing your insights and experiences. If anyone wants to continue this conversation or learn more about due diligence, investigations, and AI, feel free to connect with Christian on LinkedIn. We appreciate your time today.

Christian: Happy to be here, Rihanna. If anyone has questions or wants to discuss further, don't hesitate to reach out.

Rihanna: Thank you for listening to Fraud Talk. You can find this podcast on acfe.com, Spotify, iTunes, or wherever you listen to your podcasts. This is Rihanna Scoggins signing off.

You can listen to the full podcast below, or check it out on the ACFE website!