How to learn Artificial Intelligence: from beginner to advanced

Learn AI in five stages, in this order: 1) talk to AI well (prompt engineering); 2) use AI tools on real tasks; 3) understand the data foundation; 4) start building with coding agents such as Claude Code and Codex; 5) put a project of your own online. Skipping stages leaves gaps — each course ahead assumes the foundation of the previous ones.

See the recommended order

Where to start: the Beginner Track

This is the progression INEMA recommends for anyone starting out. It is not a random list: you learn to talk to AI, see those skills in action, build a technical foundation and only then move on to coding agents, which bring everything together in practice. All eight courses are available in English, Spanish and Portuguese: use the language links next to each one.

  1. FEP — Prompt Engineering — Learn to talk to AI: give context, ask for a format, iterate. It is the skill every other one relies on. (Português · English · Español)
  2. ATIA — Digital Opportunities with AI — See prompts working in real tools and understand what each type of AI solves. (Português · English · Español)
  3. FDB — Database Fundamentals — The technical data foundation: almost every AI project reads, stores or organizes information. (Português · English · Español)
  4. VISION — AI Audiovisual — Take AI beyond text by working with images. (Português · English · Español)
  5. Claude Code Basic — Complete Course — Your first coding agent: installation, commands, skills and MCP. (Português · English · Español)
  6. Codex Basic — Complete Codex CLI Course — The other major coding agent, to compare and choose your own workflow. (Português · English · Español)
  7. From Zero to Deploy — From your first terminal line to an AI assistant — Put a real project online, from your first terminal line to your own AI assistant. (Português · English · Español)
  8. INTELECTO — From Zero to Expert in AI — Round off the foundation by building your own AI system. (Português · English · Español)

Want to understand the role of each course in detail? See the Beginner Track guide (in Portuguese).

After the foundation: choose a track by goal

With the foundation in place, go deeper into what matters for your work. INEMA’s catalog has 310 public courses (most of them in Portuguese) organized in these tracks:

Or search the full catalog of AI courses.

How to study AI without getting lost

  • Practice on a problem of your own from day one — a report, a support routine, a spreadsheet.
  • Learn one tool at a time until it solves something real; only then move on to the next.
  • Keep what worked: prompts, instructions and mistakes become your reference material.
  • Swap "loose chat" for a system: folders, fixed instructions and memory for your agents.
  • Study in a group: asking questions and seeing other people’s projects saves months of trial and error.

Common beginner mistakes

  • Collecting courses and tools without finishing a single project.
  • Skipping prompt fundamentals and jumping straight to complex agents.
  • Asking for everything in a single message, with no context or example of the expected result.
  • Trusting the AI’s answer without checking — especially numbers, links and code.

Other INEMA guides (in Portuguese)

Learning together: the INEMA community

Learning AI alone is possible, but slow. At INEMA you get hands-on courses, open-source projects to study and a community to ask questions and keep up with what is new. Ongoing training is at INEMA.PRO, which includes the INEMA.VIP community (in Portuguese).

Frequently asked questions about learning AI

Do I need to know how to code to learn AI?

Not to get started. Prompts, ready-made tools and much of automation require no code. Programming starts to help when you reach coding agents (Claude Code, Codex) and want to put projects online — and those agents are exactly what lowers the barrier to programming.

Where should I start learning artificial intelligence?

With prompt engineering: the skill of giving context, asking for a format and iterating with AI, which every later stage depends on. Then hands-on tools, the data foundation, coding agents and, finally, a real project published online.

How long does it take to learn AI?

It depends on the goal. Using AI day to day with good results takes a few weeks of practice. Building your own agents and systems takes months of projects. What speeds it up is practicing on real problems, not piling up courses.

Are INEMA courses free?

Yes. INEMA.club courses are open and free: each course page links straight to the course itself, and the eight Beginner Track courses are available in English, Spanish and Portuguese. Ongoing training, with complete tracks, exams, certificates and the INEMA.VIP community, is part of INEMA.PRO.

What is the difference between using AI and building with AI?

Using AI means asking an assistant for answers. Building with AI means organizing context, tools and rules so that agents carry out tasks end to end. The path in this guide takes you from the first to the second.