Block 7. Agentic AI for Video

Generative AI answers a prompt once and stops; agentic AI runs a loop — it observes, reasons, acts with tools, checks the result, and repeats until a go…

A model becomes an agent when you give it three powers it does not have on its own: tools to reach the outside world, memory to carry knowledge across s…

An agent framework is the toolkit your engineers use to wire up the three agent primitives — tool use, memory, and planning — without rebuilding them fr…

A video investigator agent is an AI system that takes a plain-language question about recorded footage — "did anyone enter the loading bay after 22:00 l…

A meeting copilot agent is an AI system that sits inside a live video call, follows the conversation, and helps the people in it — surfacing the right d…

An async video review agent is an AI system that works through a large pile of recorded video on its own schedule — a back-catalog, a moderation backlog…

AgentOps is the practice of running an AI agent in production once it is built — the discipline that answers four questions you cannot ship without: can…

The previous lesson covered the three established agent frameworks — LangGraph, CrewAI, and AutoGen — but 2026 added a second wave of tools that a video…

A video archive investigator agent is a single AI system that turns a silent pile of recorded footage into something you can ask questions of in plain l…