2. AI Fundamentals and Literacy (24%)
A procurement lead sits in a vendor meeting. The vendor says the product "uses generative AI fine-tuned on your data, with an agentic layer and RAG for grounding." Everyone nods. Three months later the company discovers that "fine-tuned on your data" meant a one-time training run on last year's documents, that the "agentic layer" can issue refunds without approval, and that nobody planned for the model's accuracy to slip as products changed. The vocabulary was there; the literacy was not.
This domain is about that literacy. It is not a computer science course. It is the working vocabulary a business leader needs to ask the right questions, choose the right type of solution, and spot when a proposal is mismatched to the problem. It covers three tasks: the core concepts and terminology of AI (Task 1.1), how to select an appropriate solution type, including when not to use AI and how to keep AI under control after launch (Task 1.2), and the practical generative AI concepts (prompts, tokens, RAG, fine-tuning) that shape quality and cost (Task 1.3).
Every concept here connects back to the AI value chain from Phase 1. Training and data quality determine what the model can know; solution type determines how predictions become actions; monitoring determines whether value survives contact with a changing world; and prompting and adaptation determine whether generative AI produces useful, grounded output.
⚠️ Exam Trap: Domain 1 questions are rarely pure definitions. They usually embed a term in a scenario and ask for a decision: which solution type, which adaptation technique, which control. Knowing the definition is necessary but not sufficient; know what each concept implies for cost, risk, and fit.
