Agentic Workflow

User: “I want to sign up for the Machine Learning Bootcamp.”
LLM Function Call #1 → Information request (DB lookup): • get_course_info(course_name) • Returns details: price, start date, available seats.
Dependency: Only if seats are available does the system proceed.
LLM Function Call #2 → Action initiation: • enroll_student(student_id, course_id) • Attempts to add the student to the course roster.
Exception process: • If enrollment fails (e.g., class is full, payment method invalid), trigger a different flow: • Suggest an alternative date or course → get_alternative_courses(topic) • Or retry with a different payment option.
Follow-up Action (dependent): • Once enrollment succeeds, trigger a secondary action: • send_welcome_email(student_id, course_id) • This shows chaining of actions after success.