Master key AI use cases in finance: credit scoring, fraud detection, and automated financial reporting
This one-day training course allows finance professionals to understand the fundamental principles of Artificial Intelligence, its learning methods, and its concrete applications in the financial sector: credit scoring, fraud detection, and financial reporting automation. It combines conceptual insights, industry examples, and hands-on practice with accessible generative AI tools (such as ChatGPT, Copilot, or equivalent).
Is it for you ?
Finance professionals, financial analysts, compliance officers, and anyone interested in the application of AI in finance
Prerequisites
Basic knowledge of finance, general understanding of statistics (no programming required)
What You'll Walk Away With
- ✓ Understand the principles of AI and distinguish between symbolic AI, Machine Learning, and Deep Learning
- ✓ Identify the 3 types of learning (supervised, unsupervised, reinforcement) and their use cases in finance
- ✓ Understand the logic behind automated credit scoring and associated performance metrics
- ✓ Understand fraud detection techniques and the challenges of imbalanced classes
- ✓ Use generative AI to automate and accelerate the production of financial reports
- ✓ Situate these uses within the regulatory framework (Law 25, explainability) and build a personal action plan
At the end of the day, participants will be able to:
Training content
1 What is AI?
- Symbolic AI vs. Machine Learning vs. Deep Learning: key differences
- What AI does well, what it doesn't (prediction, hallucination, no direct access to business systems)
- Always verify critical data before distribution
2 The 3 types of learning
- Supervised learning: prediction from labeled data
- Unsupervised learning: anomaly detection, segmentation
- Reinforcement learning: optimization through feedback
3 Overview of AI use cases in finance
- Automated credit scoring: customer risk analysis
- Fraud detection: suspicious transactions, identity theft
- Cash flow forecasting (time series)
- Automated generation of summaries and financial reports
4 Discussion & Q&A
- Guided discussion: what use cases have participants already encountered?
- Identifying 2–3 business priorities for the day
5 Supervised methods for credit scoring
- Logistic regression vs. decision trees: logic and differences
- Concept of explanatory variables and probability of default
6 Measuring model performance
- Precision, Recall, F1-score: what are they for?
- Finance-specific metrics: default rate, false positives / false negatives
- Why the cost of a false negative is not the same as a false positive in credit
7 Guided workshop – Reading a scoring case
- Study of a simplified numerical credit scoring example
- Collective interpretation of results and their limitations
- Discussion: where does the model end and human judgment begin?
8 Fraud detection techniques
- Typology of fraud: credit card, wire transfers, identity theft
- Impact of fraud on financial organizations
- Unsupervised methods: Isolation Forest, Local Outlier Factor (LOF)
- The challenge of imbalanced classes (very few real fraud cases)
9 Generative AI for financial reporting
- Principles of the RCTF method (Role – Context – Task – Format) to build a useful prompt
- Use cases: summarizing a results table, commenting on budget variances, drafting risk notes
- Best practices: provide figures in the context, verify numbers before distribution
10 Practical workshop – Generating a reporting commentary
- Starting from a fictional Actual vs. Budget table, drafting a prompt using the RCTF method
- Generating a summary in a few lines, professional tone
- Iteration: rephrasing, enriching, adapting tone and format
- Collective correction and best verification practices
11 Regulation and transparency of AI models
- Reminder of Law 25 principles applied to scoring and detection models
- Why model explainability is becoming a business and regulatory requirement
12 My action plan
- Identify 1 priority use case to test in your activities (scoring, fraud, or reporting)
- Choose 1 generative AI tool to experiment with the following week
- Tip: block 15 minutes per week to test a new prompt
13 Review & wrap-up
- Recap of key takeaways from the day
- Open Q&A
- Handout of materials and prompt examples
📌 Practical information
Our training sessions are offered in Montreal or Quebec City, in person or in a virtual classroom. Dates and locations are specified when you select your session below. If you have any questions, check out our FAQ.
