This hands-on course introduces accountants, auditors, and finance professionals to agentic AI by simulating autonomous behavior using the basic version of ChatGPT. Participants will use role-play, prompt engineering, and guided interaction loops to create agents that simulate planning, decision-making, and communication workflows. No coding or advanced access is required. Participants will practice designing task-specific prompts, simulating agent autonomy using flipped interaction patterns, and evaluating agent output for risk and effectiveness. This course makes agentic AI accessible to all learners through practical, scenario-based exercises using tools they already have.
Learning Objectives
- Understand the building blocks of agentic behavior using basic ChatGPT
- Create and simulate task-specific agents using prompt engineering techniques
- Practice flipped interaction patterns that showcase autonomous agent workflows
- Evaluate agent responses for bias, limitations, and control weaknesses
- Apply ethical and professional guidelines to simulated agent deployments
Major Topics
- Agentic simulation using prompt design and structured interaction
- Flipped agent/user roles in ChatGPT (autonomy emulation)
- Use case walkthroughs: audit Q&A bot, tax helper, financial planning agent
- Risk scenarios: hallucinations, scope creep, prompt hacking
- Governance, professional skepticism, and documentation best practices
