Fraud & Financial Crime Detection with Artificial Intelligence (AI)

Session 1 – Setting the Stage

• Introduction to Financial Crime & AI
• The landscape of financial crime: fraud typologies (transaction fraud, insider trading, money laundering, cyber-enabled crime)
• Regulatory drivers: AML, KYC, FATF standards, sanctions screening
• Why AI/ML is relevant: limitations of rule-based systems, explosion of data
• Overview of AI techniques used in fraud detection: supervised, unsupervised, and hybrid models
• Exercise: short case analysis – “why rules failed” in a real fraud case

Session 2 – Data Foundations

• Data for Fraud Detection
• Types of data: transaction records, client profiles, network/graph data, unstructured text (emails, chat logs)
• Data challenges: imbalance, noise, false positives/negatives
• Feature engineering for fraud detection (e.g., unusual transaction patterns, velocity, link analysis)
• Role of NLP in compliance (screening suspicious narratives, contracts, chatrooms)
• Demo: walk through a Jupyter Notebook showing basic anomaly detection on a synthetic transaction dataset

Session 3 – Core AI Techniques

• Machine Learning Methods for Fraud & Financial Crime
• Supervised learning: logistic regression, random forests, gradient boosting
• Unsupervised learning: clustering, autoencoders, anomaly detection
• Graph-based AI: detecting collusion rings and layered transactions
• LLMs & Generative AI: assisting investigators (summarising suspicious activity reports, extracting red flags from documents)
• Exercise: hands-on guided notebook – train a simple classifier on fraud vs non-fraud data

Session 4 – Practical Applications

• Industry Use Cases
• Banking & payments: transaction monitoring, card fraud
• Insurance: claims fraud
• Trading & markets: detecting manipulation, insider trading
• AML: beneficial ownership, ultimate parent tracing with NLP and graph search
• Integrating AI into compliance workflows: alert triage, SAR drafting, regulator engagement
• Case study: how a bank applied AI to reduce false positives in AML monitoring by 30%

Session 5 – Governance, Ethics, and the Future

• Trustworthy AI in Financial Crime Detection
• Regulatory expectations for explainability (FCA, ECB, MAS guidance)
• Bias, fairness, and the danger of over-reliance on “black box” models
• Human-in-the-loop models: investigator + AI collaboration
• Future trends: federated learning for cross-institution fraud detection, privacy-preserving analytics, generative AI for red team testing

Key takeaways

By the end of the day, participants will be able to:

• Understand and identify common fraud and financial crime patterns with AI and explain why traditional detection methods often fall short
• Prepare and interpret financial datasets for fraud analytics, including dealing with class imbalance and suspicious transaction features
• Apply core AI and machine learning techniques — both supervised and unsupervised — to detect anomalies and suspicious behaviours
• Evaluate fraud detection models using the right metrics, to reflect operational realities
• Explain and communicate model outputs in a way that meets governance and regulatory expectations, supporting trust and adoption in compliance teams

The course trainer is a financial educator and technologist with over 25 years of international experience at the intersection of front-office finance, data science, and artificial intelligence. He began his career in banking technology at UBS and JP Morgan, where he led projects redesigning trading platforms used on global derivatives desks. Building on a First-Class degree in cognitive psychology with a dissertation in artificial intelligence, and the Certificate in Quantitative Finance (CQF), he has since trained thousands of professionals worldwide in derivatives, risk management, Python, AI, and machine learning.

His work spans teaching at London Business School and Cambridge Judge Business School, to delivering programmes for Goldman Sachs, Morgan Stanley, Credit Suisse, HSBC, and sovereign wealth funds. He has designed AI-based training platforms in partnership with the CFA Institute and is actively developing applied AI systems for both legal compliance and trading pattern recognition.

Recognised for making complex topics clear, rigorous, and practical, he brings together deep financial markets expertise and hands-on AI experience. This unique combination positions him to show participants not only how AI detects fraud and financial crime, but also how to apply these techniques effectively, responsibly, and in line with regulatory expectations.

Upon completion of this Fraud & Financial Crime Detection with AI course, participants will be able to:

• Recognise key fraud and financial crime patterns — understand how fraudsters exploit systems and why traditional rule-based detection often fails.
• Understand the role of AI in modern detection — gain clarity on which machine learning and AI methods are most effective for transaction monitoring, anomaly detection, and compliance.
• Work with real-world style data — learn how to prepare and interpret transaction and client datasets for fraud analytics, including handling class imbalance and noisy signals.
• Apply core AI techniques — develop hands-on familiarity with supervised, unsupervised, and anomaly detection models to identify suspicious behaviours.
• Evaluate models with fraud-relevant metrics — move beyond accuracy to focus on precision-recall, false positives/negatives, and cost-sensitive trade-offs.
• Explore explainability and governance — appreciate how to make AI models transparent, auditable, and regulator-ready in financial crime contexts.

This course is designed for finance and compliance professionals who need to spot fraud faster, reduce false positives, and understand how AI can be safely applied in a regulated environment. The target audience is:

• Risk, compliance, and audit professionals — who need to understand how AI strengthens detection frameworks and reduces false positives
• Financial crime investigators and analysts — who want to see how machine learning can uncover hidden patterns and networks of suspicious activity
• Data scientists and technologists in financial services — who are interested in applying AI methods to fraud detection but need domain-specific context
• Project managers and innovation leads — who oversee AI or Regulatory Tech initiatives and want to translate technical methods into business value
• Regulators, supervisors, and consultants — who wish to understand current AI practices in fraud and financial crime to engage with institutions more effectively

‘Fraud & Financial Crime Detection with AI’ is a one-day practical course that equips finance and compliance professionals with the knowledge and tools to harness artificial intelligence in combating fraud. Through real-world case studies, hands-on demonstrations, and clear explanations, participants will learn how to prepare data, apply core machine learning techniques, and evaluate models using fraud-relevant metrics. The course also addresses explainability, governance, and regulatory expectations, ensuring that AI solutions are not only effective but also trusted and compliant. By the end of the day, attendees will understand how to integrate AI into financial crime detection workflows to reduce false positives, uncover hidden risks, and strengthen their organisation’s defences.