Agentic AI for the audiovisual industry
Stop chatting with AI and start delegating: agents that organize, classify and prep your footage while you edit.
A chatbot answers questions; an agent gets work done. This course breaks down the architecture of AI agents — the harness, the agentic loop, skills, MCP connectors — and lands it in real audiovisual workflows: automated ingest, metadata-driven classification, computer-vision tagging and sidecars ready for Premiere or DaVinci Resolve. No coding required: you will learn to direct agents with professional judgement, with the permissions, checkpoints and safety controls a real production demands.
By the end of the course you'll know how to
- Tell a chatbot from an agent with confidence, and know when to delegate each task.
- Understand the full agentic harness: models, actions, connectors, skills and triggers.
- Read and direct the agentic loop: plan, context, execution and goal check.
- Configure the agent's identity and memory with Agents.md, Claude.md and Memory.md.
- Plug the agent into your ecosystem with reusable skills and the MCP protocol.
- Apply permissions, checkpoints and supervision to delegate without babysitting.
- Automate a real ingest: classification by Camera ID, Reel, date and frame rate with sidecars for Premiere and Resolve.
- Generate descriptive metadata with vision: searchable libraries without hundreds of hours of manual review.
Is it for you?
This course is for you if…
- Editors, colorists, DITs and post producers who want to automate repetitive work.
- Producers and technical leads evaluating agents for their pipeline.
- Creative profiles with no coding background: this is about judgement, not code.
It's not for you if…
- Anyone looking for an LLM programming or machine-learning-from-scratch course.
- Anyone after image or video generation prompts: here agents do the work, they don't generate shots.
Dubbing into your language in production. For now, this course has Spanish audio with subtitles in your language.
Curriculum
1. Introduction to the agentic paradigm2lessons
- From chatbot to agent
- Definition and the agentic harness
2. Anatomy of the agentic loop3lessons
- The think-and-act cycle
- Configuring with Agents.md and Claude.md
- Dynamic memory: Memory.md and projects
3. Skills and the connection ecosystem3lessons
- Reusable skills
- Connectors and the Model Context Protocol (MCP)
- Tools in use: Cowork, Codex and browser use
4. Safety, control and supervision3lessons
- Agentic risks and the babysitting effect
- Checkpoints and control points
- Permission management and human approval
5. Hands-on: agentic AI in the audiovisual workflow3lessons
- Use case: automated ingest
- Automated descriptive metadata
- Conclusion and next steps
Frequently asked questions
How long is the course?
This course has 2 hours of video across 14 lessons and 5 sections. Watch it whenever you want, at your own pace — no mandatory live classes.
Does access expire?
No. You pay once and access is unlimited: you can watch and rewatch the videos as many times as you want, with no subscription and no deadline.
Is there a diploma? How is it verified?
Yes. When you finish the course and pass its final exam you get a diploma with a unique code that anyone can verify publicly at jmochon.training/certificado, plus a digital badge (Open Badge) for your LinkedIn profile.
What languages is it available in?
This course’s audio is in Spanish and the available subtitles are: Spanish, English, Portuguese.
What level do I need to start this course?
This is a beginner-level course: you don’t need any prior experience with the software, it’s explained from scratch.
Can I get an invoice under my company’s name?
Yes. Payment is processed securely with Stripe, and after purchase you can request an invoice with your company’s billing details.
What is the AI tutor?
An assistant available 24/7 inside the course: ask it any question and it answers citing the exact lesson and minute of the video where it’s explained, at no extra cost.