# Lesson 2 ยท The domain everyone wants to skip This is the only thing you record. The slides run on their own. Pace measured on your English take: 168 words per minute. These 890 words are about 5:17. --- Welcome to Domain 1: AI fundamentals and literacy. This is the first domain on the exam, and it's the one most people skip. It looks like vocabulary, and most of us assume we already have the vocabulary. Let me show you why that assumption is expensive. Picture three different meetings. In all three, someone says exactly the same sentence: we're using artificial intelligence. In the first one, they mean a generative assistant. That's paid for per token consumed, and the risk is that it makes things up. In the second one, they mean a model that estimates something: how much, when, or how likely. That's paid for per prediction, and the risk is that it decays over time. And in the third one, they mean a rules script. Which is not artificial intelligence. It's built once, and the risk is that it goes stale. Three different things. Three different costs. Three different risks. And the same sentence. I want to stop on the third case, because it's the most common one and the most expensive. Notice one detail, the one that explains everything: the person saying that sentence in the meeting is not the person who wrote those rules. Somebody else wrote them, and they could write them because they knew every case in advance. If this happens, then that. What the person in the meeting sees is something else. They see a system that resolves on its own, with nobody stepping in. And today, anything that runs on its own gets called artificial intelligence. That's the difference you need to be able to say out loud. A rule runs decisions someone already made. A model makes decisions nobody made before, by learning from examples. From the outside they look alike, because both run on their own. On the inside they have nothing in common, and you find that out when the invoice arrives and when something goes wrong. That's why the misunderstanding costs money: it ends in an AI budget approved for something that isn't AI, and AI governance demanded of something that doesn't need it. So this domain isn't vocabulary. It's the difference between approving a budget that makes sense and approving one that doesn't. Now, why is this yours to solve? Because you, as an AI business strategist, you evaluate, champion and scale AI initiatives across the organization. You work with the technical teams, but you don't build the models. Your value is something else: making sure that what gets decided in the executive committee is what actually gets built by the technical teams. Without that shared language, governance gets distorted and the budget ends up misallocated. With it, every initiative in the company becomes comparable, which is what lets you decide which one goes first. And here's something that surprised me. This exam expects you to ask three questions of every initiative. Whether artificial intelligence is the right tool for this problem. What it will cost and what it will return. And how do we govern it so it keeps working. Those are exactly the three steps of the route I showed you in the previous video. Identify where artificial intelligence should act. Select the case that's worth it. And implement without the stress, with governance as the floor. The exam asks what the route answers. I didn't design it for this exam, but it fits, and that's why we'll walk all four domains with it. This domain has three tasks. The first one is concepts and terminology: the vocabulary itself. The second one is choosing the type of solution: when a rule is enough, when you actually need artificial intelligence, what an agent is, and how you monitor what's already running. And the third one is working with generative artificial intelligence: prompts, tokens, and the techniques for improving answers. They're chained: each one needs the one before it. Without the vocabulary you won't choose the right type of solution, and without choosing well you won't get much out of the generative part. One last thing before we start. Throughout this domain we're going to use a real case, not an invented one. It's called AgentSquad, and it's a platform where artificial intelligence agents run a company's processes. I picked it because I built it, which means that when we talk about why a model that was working starts to fail, what it costs to keep it running, or who answers when it gets something wrong, I'll be telling you what actually happened, not what would happen. Today there are two ideas on the table, and a decision to make. The first one: an assistant that answers what can be done and who has to approve it, according to the operating rules. The second one: the system deciding on its own which model gets each job. Are they the same thing? Are both of them artificial intelligence? Which one goes first, and what does each one cost? By the end of this domain you'll be able to answer that out loud, in front of whoever needs to hear it. And to get there we start with the vocabulary. That's what the next lesson is about: what an AI business strategist actually does, and the handful of words you need to understand the technical teams and, with those same words, whoever approves the budget. See you there.