What this article is about
The most common way to get lost learning AI is to treat every class as a one-off tool demo: slides today, transcripts tomorrow, and still no idea what comes next. The point of a course map is to answer "what do I learn next": to connect each class to a capability node, so the things you learn link up with each other. This article lays out the map for my entire course lineup.
· People who have taken a few AI classes and want to know where they stand now and where to go next
· People who want to move systematically from getting started with tools to building a personal work system
· Teachers and knowledge workers designing their own curriculum, who can borrow the logic of this map
· The structure of the whole course map: a foundation layer, applying upward, and digging deeper downward
· The two layers of each of the three deep-dive lines, and the destination where they converge
· The logic that connects the courses: each step naturally raises the question that leads to the next
The shape of the map: one foundation, two directions
The whole map is a vertical structure. In the middle is a foundation layer, the starting point for everyone: the Agent Onboarding basics course, which covers the most fundamental parts of setting up an AI Agent working environment: software installation, environment setup, and core concepts. After this class you walk away with a basic Agent environment that actually runs, something I call the "AI Manager Lite" version. It is the prerequisite for every route that follows.
From the foundation layer, the map opens in two directions, one going up and one going down:
Take ready-made skill packages and workflows and use them directly to solve the problem in front of you fast, without deep study. Things like the meeting-notes workflow or the sustainability-report workflow run the moment you pick them up. Most of them are available on the Skills download page.
Through courses and coaching over the long term, research the areas you are good at and build your own system. Slow, but it compounds.
The two paths do not conflict. If you are strong at knowledge management but weak at sustainability reporting, dig down into tacit knowledge distillation while applying the ready-made sustainability-report workflow going up. Dig deep where you are strong, borrow others' tools where you are not, and take both at once.
Digging deep: three parallel lines, two layers each
The downward direction has three parallel lines, each extending straight from the foundation layer and going one layer deeper:
Organize your data the AI way, so your knowledge base can be read and used by AI. Beyond learning the tools, the data-organization class is really about making your data usable by AI.
Turn one-off experience into reusable knowledge assets by first digging out the things you "can do but can't explain."
Understand semantic space and word association, and learn to navigate AI's probabilistic world. Beyond sentence patterns, the prompt class is really about stating a task clearly.
Upgrade the AI Manager from the Lite version to the full version and build a personal management system.
Dig from emotional entry points down to deeper judgment criteria, and build your own thinking framework.
Build a controlled vocabulary across four dimensions (object, content, context, and type) so AI understands your classification logic precisely.
The three lines all converge on the same destination: an Agent-powered solo-company operations team. Use a team of AI Agents to run the operations of a solo company, where everything you learned earlier (data organization, tacit knowledge, semantic engineering, and the management system) all ties together.
How the courses connect: let the next question grow on its own
The courses on this map share one design principle: at the end of each step, learners naturally arrive at the question for the next step, with no need to push.
The free intro talk "Word Association and Prompt Design" centers on one idea: communication goes both ways, in how you speak and how AI hears you. People who finish it usually ask, "Now that I can design a good prompt, what can I do with it beyond pasting into a chat box?" That connects to the next set.
"Attention Mechanisms and Context Engineering" answers that question, covering how AI remembers and how you manage it: AI has no memory, only a window. It walks through the containers (the chat box, project mode, skill packages) and how each differs and when to use it. People who finish it ask, "I know how to use the containers now, but what matters most to put inside them?"
The answer is tacit knowledge distillation: the raw material AI runs on is your knowledge, and the most valuable pieces are usually the ones you can do but can't clearly explain. The workshop walks you through actually digging one or two layers, and then splits into two directions: dig deeper into the underlying logic and judgment criteria, or expand horizontally into rule-base design. A knowledge base lets AI empower your brain (understanding your expertise); a rule base lets AI empower your hands and feet (executing for you). String the two together and you move from "AI understands you" to "AI does things for you."
How to use this map: find where you are first
The first step in reading a map is locating yourself. Work out which stage you are at now, and the next class becomes part of a route rather than a standalone lesson.
- No Agent environment yet: you are before the foundation layer, so set up the environment first.
- Can use tools but your data is a mess: take the left line and start with data organization, so your knowledge base becomes usable by AI.
- Have the data but can't articulate your own expertise: take the middle line, tacit knowledge distillation.
- Your instructions keep getting misread by AI: take the right line, semantic engineering.
- Something urgent to solve right now: go up and apply a ready-made workflow directly, while still walking the other path.
Once you know where you stand, each class connects to your long-term capability, and learning shifts from collecting tools to accumulating a system. If you want to start from your own end first, Use AI to build a trackable learning map is the personal-scale starting point for this course map.