Microsoft · AB-731
Generative AI foundations
AB-731 · Lesson 1 of 8 · 3:11
YouTube loads only after you press play. Watch on YouTube ↗
In this lesson
Prepare for the AB-731 exam (Microsoft Certified: AI Transformation Leader). In this lesson: what generative AI is compared with traditional AI, large, small, and multimodal models, choosing between pretrained and fine-tuned models, what drives costs (a per-user subscription for Microsoft 365 Copilot, tokens for models in Microsoft Foundry) and how to measure return on investment, and the challenges: fabrications, reliability, bias, privacy, and explainability. Includes three original practice questions.
Preview the topics
- A law firm wants a model that uses its own contract terminology more accurately than a general model, and it accepts the extra cost. Which approach fits?
- A team built an assistant on a model in Foundry. Its monthly bill doubled after users started pasting long documents into every prompt. What drives the increase?
- An assistant answers a question about a company policy with a confident, detailed rule that does not exist. What is this called?