Skip to content
AI AutomationGet Pro

Menu

Taught by Claude Certified AI System Architects

The AI agents course.

Build and train custom agents that run your business processes.

Build a custom AI agent for a specific business process, train it to do the job your way, and run it on guardrails: securely, reliably and at a known cost. On schedules, on events and around the clock. Platform agnostic.

Opening soonGet it with AI Automation ProStart with AI Automation

63 of 200 places left at today's prices. Prices go up after the first 200 members.

You build
Custom AI agents and systems
Approach
Platform agnostic
Course
5 acts, 28 episodes
Autonomous AI Automation: the course's name on a chip at the centre of a circuit board

What a trained agent does

Five steps of the process, handed to your agent.

A custom agent runs a business process end to end, inside the guardrails you set.

  • 01

    Picks up the work

    It starts on a schedule or an event, without anyone asking.

  • 02

    Follows your procedure

    Trained with skills, it does the job your way, every time.

  • 03

    Moves the information

    Connected to your tools, it reads the inputs and writes the results.

  • 04

    Checks the result

    Guardrails and checks run before anything leaves the process.

  • 05

    Brings you the decisions

    It stops at the points you choose and asks you to approve.

With a custom agent

The work runs. You check the results.

One example day: jobs that start on a schedule, on an event and overnight.

How an agent works

The loop that does the work.

An agent takes a goal, decides the next step, acts with its tools, checks the result, and repeats until the job is done or it needs a person.

01Observe. Decide. Act.The whole cycle is the agent loop, continuing toward the goal one turn at a time.
02The model decides. The harness acts.Context goes to the model, instructions come back, and the harness runs the tool.
03Every turn is a new request.The loop continues with the results of the last action, until the job is done.

In the course's own words, from Episode 05, How an Agent Works.

The path

From your first agent to autonomous systems.

Four acts build the skills. In the fifth, you choose what to build.

  1. Act 1Understand AI agents
  2. Act 2Build a custom AI agent
  3. Act 3Automate the work
  4. Act 4Set up an autonomous AI agent
  5. Act 5Choose your projects

What you build

Custom AI agents you can rely on.

  • Custom AI agentsBuilt for one business process and connected to the tools it runs on.
  • Trained with skillsYour procedures, written once and executed the same way every time.
  • Run on guardrailsPermissions, checks and a known cost per task, so it works securely.
  • Running on their ownOn schedules, on events and always on, with you approving what matters.

Curriculum

What happens in each act.

28 episodes. Acts 1 to 4 build the skills on one real job; in Act 5 you choose the projects that fit your business.

  1. Act 1 · 7 episodes

    Understand AI agents

    What an AI agent is, the harness that runs around the model, the five layers up to an autonomous agent, local or cloud models, the agent loop, and the AI Implementation Map you plan every build with.

    Episode 01: an agent takes a goal, decides, acts and checks the results, until the job is done or it needs you
    Episode 01What Is an AI Agent?

    You leave this act withA plan for your agent, mapped on the AI Implementation Map.

    1. 00What You'll Build
    2. 01What Is an AI Agent?
    3. 02What Is an AI Harness?
    4. 03The Five Layers of AI Harness
    5. 04AI Models: Local or Cloud?
    6. 05How an Agent Works
    7. 06The AI Implementation Map
  2. Act 2 · 9 episodes

    Build a custom AI agent

    Set up your first local agent, connect it to your apps and to GitHub, and run your first real job: a client call becomes a project brief and owned tasks on your board. Then turn that procedure into a skill, learn what each run costs, set your guardrails, and see how Luka edits video with AI.

    Episode 07: Claude Code running in a project folder, with auto mode explained on screen
    Episode 07Set Up Your First Local AI Agent

    You leave this act withA custom agent on your machine, trained with a skill, on guardrails, at a known cost per run.

    1. 07Set Up Your First Local AI Agent
    2. 08How Agents Connect
    3. 09Connect Your Agent to GitHub
    4. 10Your First Job
    5. 11What Is a Skill?
    6. 12Turn Your Procedure Into a Skill
    7. 13What It Costs
    8. 14Set Your Guardrails
    9. 15How I Edit Video With AI
  3. Act 3 · 3 episodes

    Automate the work

    Put the job on a schedule, trigger it on an event, and learn when a plain workflow beats an agent.

    Screenshot to addThe job set as a routine at 7:00, and its first run finished
    Episode 16Run It On a Schedule

    You leave this act withYour agent's job running on a schedule and on events, without anyone starting it.

    1. 16Run It On a Schedule
    2. 17Trigger It On an Event
    3. 18Workflow or Agent?
  4. Act 4 · 6 episodes

    Set up an autonomous AI agent

    Set up an always-on machine, install OpenClaw, connect and configure it, connect your apps, and let it take its first job alone, ready for your projects.

    Screenshot to addThe always-on machine with OpenClaw running on it
    Episode 19Set Up Your Always-On Machine

    You leave this act withAn always-on agent on its own machine, connected to your apps and ready for projects.

    1. 19Set Up Your Always-On Machine
    2. 20Install OpenClaw
    3. 21Connect and Configure It
    4. 22Connect Your Apps
    5. 23Its First Job Alone
    6. 24Get It Ready for Projects
  5. Act 5

    Choose your projects

    Choose what you build next. Each project is a custom agent for a real business process; build one, several, or bring a process of your own.

    You leave this act withThe agents you chose, running the business processes you picked.

    Choose what you build

    • Content Engine
    • Video Editor
    • Research Analyst
    • Outreach Rep
    See the projects
Luka Madzarac opening Episode 01, What Is an AI Agent?Luka MadzaracClaude Certified AI System Architect

Your instructor

Taught by Luka Madzarac.

Luka teaches the course on camera. He spent a decade producing brand film, motion design and web, then built the agent system his own studio runs on. The course builds that kind of system with you: custom agents, trained to be reliable and run real work.

  • Runs his own studio on AI agents
  • Claude Certified AI System Architect
  • Shows every step on screen

Act 5: choose your projects

Pick what you build.

Every business runs different processes, so in Act 5 you choose: build one of these agents, several, or bring a process of your own.

  • Choice 1

    Content Engine

      • Your notes
      • Your recordings
    1. Agent drafts
    2. You approve

    Turns your notes and recordings into a week of posts for your channels, drafted for your approval.

    Runs
    Every week
    Guardrail
    Nothing posts until you approve it
  • Choice 2

    Video Editor

    1. Raw footage
    2. Agent cuts
    3. Your review
    4. It ships

    Until you approve

    Takes raw footage to a first cut the way you edit, ready for your review.

    Runs
    When new footage lands
    Guardrail
    You review every cut before it ships
  • Choice 3

    Research Analyst

    1. Every Monday
    2. Agent researches
    3. The brief

    Again next week

    Delivers a competitive research brief on your market every Monday, without anyone starting it.

    Runs
    Every Monday morning
    Guardrail
    Works only from the sources you set
  • Choice 4

    Outreach Rep

    1. Prospects
    2. Agent writes
    3. You approve
    4. Sent

    Finds the right prospects, writes a personal first message for each, and sends it on your approval.

    Runs
    Every weekday
    Guardrail
    Nothing sends until you approve it

Or bring a process of your own

Pick a process from your business and build the agent for it, with a mentor beside you from the plan to the guardrails.

Apply for mentorship

Who it is for

Built for business processes.

For owners, operators and builders implementing AI in a business: agents trained on real processes, running securely on guardrails.

  • You run processes that repeat

    Reports, follow-ups, client work: the jobs a custom agent executes.

  • You want systems, not prompts

    Agents with skills, guardrails and known costs, built to be relied on.

  • You want what lasts

    Platform agnostic: the principles hold as models and tools change.

Where it sits

From workflows to agents that work on their own.

AI Automation teaches level one, workflows that run a process. This course takes you through the next two.

  1. Level 2

    AI agents

    An agent with your tools and your procedures, taking on a real job step by step.

    Autonomous AI AutomationThis course

  2. Level 3

    Autonomous AI agents

    Agents that run on their own, on a schedule, on events and overnight, while you check the results.

    Autonomous AI AutomationThis course

Your mentors

Two mentors, both courses.

We mentor both courses ourselves. Mentorship is by application: you build your own project, and whichever of us it needs guides you from the idea to something that runs.

Anthony Lee, program mentorMentor

Anthony Lee

AI SEO researcher, tool builder, agency partner

From Amazon algorithm work to AI search and the tooling around it. He runs the build-along sessions and mentors the final project, from the idea to something that runs.

TeachesAI Automation

Luka Madzarac, Claude Certified AI System ArchitectMentor

Luka Madzarac

Claude Certified AI System Architect

A decade in brand film, motion design and web, then he built the agent system his own studio runs on. He mentors the creative side: production, web builds, and the agents behind them.

TeachesAutonomous AI Automation

  • It's one of the most beautiful courses I've ever experienced. The graphics The animation The resources

    Dave KatagueMultidisciplinary Creative & AI Strategist · Course student

    Dave Katague on LinkedIn (opens in a new tab)
  • Value-packed course and next-level community. Luka and Anthony go above and beyond, constantly sharing real-time tips, insights, and blueprints that are immediately useful. I'm genuinely grateful to have been part of this experience. Wishing the whole team continued success!

    Rio GonzalesHead of Operations · Course student

  • Very good program. Totally worth it the investment. It requires a lot of focus and dedication but the outcomes are very good!

    Giuseppe MiloProduct & Growth Manager · Course student

    Giuseppe Milo on LinkedIn (opens in a new tab)
  • I'd recommend this program to anyone whether they're looking to streamline their personal tasks or scale automation inside of their own business, because this program truly covers everything you need to become an automation master.

    Jamar JonesMarketing Agency Founder & CEO · Course student

  • It is an excellent community for learning AI workflow automations! The course content is high quality alongside support from the community.

    Wajeeh QSolo Entrepreneur · Course student

  • Great minds sharing innovative information as it develops in the fast developing industry of AI

    Claude Edwin TheriaultSolo Entrepreneur · Course student

    Claude Edwin Theriault on LinkedIn (opens in a new tab)

FAQ

Questions before you start.

Do I need to code?

No. Every step is shown on screen, and terminals, agents and skills come up only after you have seen what each one is for.

Which tools and platforms do you teach?

Platform agnostic. Models and tools change every few months; how an agent works, how to give it skills, guardrails and a schedule, and how to run it reliably do not. You learn the parts that last, shown today on Claude Code.

Can I choose what I build?

Yes. Acts 1 to 4 build the skills on one real job. In Act 5 you choose your projects: a Content Engine, a Video Editor, a Research Analyst, an Outreach Rep, or a process from your own business.

When does it open?

Soon. Until then, AI Automation is open now and is step 1 of the same path.

How is it different from AI Automation?

AI Automation teaches you to build AI workflows for business processes on no-code tools. Autonomous AI Automation teaches you to build custom AI agents, train them to execute reliably and cost-effectively, and run them on their own.

RecommendedEvery course, start to finish

AI Automation Pro

The platform and every course in one subscription, from your first AI workflow to agents that run on their own.

  • AI Automation, open now
  • Autonomous AI Automation, the day it opens
  • The platform, its app and the community
  • Weekly live workshops and the blueprint library

$99/month

63 of 200 places left at $99. The price goes up after the first 200 members.

Get AI Automation Pro

One subscription. Cancel any time.

Opening soon. Start with level one.

AI Automation teaches the workflows that run a business process, the step before agents. It is open now.

Get Pro