1–2 weeks
Complete the fundamentals and prompt engineering modules. You will be able to delegate simple writing and research tasks confidently.
Learning roadmap
Learn AI with a practical builder's path
A practical curriculum for turning AI curiosity into usable skill. Each module connects to a guide, exercise, checklist, or workflow you can apply to real founder, developer, or marketing work.
Time
How long does it take
3–6 weeks
Add productivity workflows, tool selection, and RAG basics. You can build repeatable personal or team systems.
2–4 months
Cover MCP, agents, automation, and evaluation. You can design and ship internal AI tools with proper governance.
Pitfalls
Common mistakes
Chasing every new model
Most workflows do not need the latest benchmark winner; they need clear prompts and review gates.
Skipping evaluation
If you cannot define good output, you cannot tell whether AI is helping.
Ignoring data risk
Pasting sensitive documents into the wrong tool can create compliance problems.
Expecting magic
AI amplifies good workflows and exposes bad ones. Start with one real task, not a broad transformation.
Roadmap
How to use this roadmap
Work through the modules in order if you are new to AI, or jump to the section that matches your current bottleneck. Each module includes a practice exercise; completing it is more valuable than reading passively. Use the linked guides and checklists to go deeper, and return to the evaluation module whenever you add a new tool or workflow.
Roadmap
Eight modules from AI basics to governance
AI fundamentals
Understand models, tokens, context windows, hallucinations, privacy, and where AI is useful.
- For: beginners and managers
- Practice: rewrite one task for AI
- Time: 2 hours
Prompt engineering
Learn task framing, examples, constraints, output formats, evaluation, and prompt reuse.
- For: anyone producing AI output
- Practice: build a reusable prompt
- Time: 3 hours
Productivity workflows
Use AI for writing, research, summaries, planning, and reusable daily work systems.
- For: operators and founders
- Practice: document one workflow
- Time: 2 hours
AI tools by role
Map tools to writing, coding, research, automation, customer support, and operations.
- For: teams choosing tools
- Practice: score three tools
- Time: 2 hours
RAG and knowledge systems
Use retrieval, embeddings, source-backed answers, and evaluation habits for reliable outputs.
- For: builders and knowledge teams
- Practice: design a RAG flow
- Time: 4 hours
MCP and integrations
Connect AI systems with tools, files, databases, APIs, and controlled business workflows.
- For: technical teams
- Practice: map tool permissions
- Time: 3 hours
Agents and automation
Design agent workflows with planning, review gates, logs, human approvals, and testing.
- For: builders and ops teams
- Practice: add review gates
- Time: 4 hours
Evaluation and governance
Create acceptance criteria, privacy boundaries, audit logs, and team operating rules.
- For: leaders and teams
- Practice: define eval criteria
- Time: 3 hours
Choose your path
Recommended tracks by role
AI stack for small teams
Learn tool selection, automation, and adoption checkpoints.
View toolsContent and campaign workflows
Use prompts, repurposing, research, and review gates.
View workflowsCoding agents and RAG systems
Use agents for scoped repo work and design source-backed systems.
Read guideSales and GTM workflows
Use AI for prospecting, outreach, CRM hygiene, and launch planning.
Read hubWatch: a one-hour introduction to large language models
Andrej Karpathy covers what LLMs are, how they reason, and how to use them for research, writing, and coding.
Levels
A level-by-level learning roadmap
Use these four levels to pace yourself. Each level builds on the previous one, and you can stop at the level that matches your current role.
AI literacy
Learn how models, tokens, context windows, and hallucinations work. Practice rewriting one daily task for an LLM.
Start herePrompting and workflows
Build reusable prompts, run weekly research scans, and document one repeatable workflow with clear review steps.
Get checklistTools and RAG
Score tools by workflow fit and risk. Design a RAG flow that grounds answers in your documents and knowledge base.
Learn RAGAgents and governance
Build agent workflows with review gates, tool permissions, and evaluation. Add privacy rules and audit habits.
Explore agentsFirst reads
Recommended first guides
How to learn AI in 2026
The full learning path from fundamentals to agentic workflows and governance.
Read guidePrompt engineering checklist
A practical checklist for clearer, more consistent AI outputs across any model.
Read checklistBest AI tools for beginners
Choose your first AI writing, research, coding, and productivity tools.
Compare toolsWhat is RAG?
Understand retrieval-augmented generation and how to ground AI in real sources.
Read guideAI agents explained
Plan safe, useful agent workflows with review gates and clear boundaries.
Read guideFAQ
How long does the full roadmap take?
Most people reach Level 2 in two to four weeks, Level 3 in one to two months, and Level 4 in three to six months depending on technical depth.
Can I skip the fundamentals?
You can skip ahead if you already understand tokens, context windows, and hallucinations, but revisit them when output quality or safety becomes a bottleneck.
Which level is right for founders?
Founders should aim for Level 2 quickly, then use Level 3 for tool selection and RAG decisions. Level 4 matters once you are building or buying agentic systems.
Do I need a paid subscription to start?
No. Start with free tiers of ChatGPT, Claude, Gemini, or Perplexity. Pay only after a tool proves value on a real weekly workflow.
How do I know I am making progress?
Track one metric: time saved, output quality, or decisions improved. When a workflow feels repeatable and reviewable, you have reached the next level.