77 IA Course

Free course · Updated October 2026

Understand AI.
Then put it to work.

A practical course for business owners and teams: what AI really is, how to use it well at work and, if you want to go further, how to build with Claude Code and agents. Written by someone who uses it every day to run a real company.

Start with module 1 See the outline
modules
21modules
parts
4parts
of reading in total
6 hof reading in total
published
10/21published

How every module is built

01

The concept

Explained simply, with business examples and no filler.

02

Try it now

A hands-on exercise of about 15 minutes.

03

Prompts to copy

Ready-to-paste instructions for Claude, ChatGPT or Gemini.

04

Real case

How we use it at 77Rentals to run the company.

05

Mini quiz

Four questions to check it stuck.

The outline

21 modules, 4 parts

Four parts, in order. The first two are for everyone; the third and fourth are the builders’ track, but anyone can read them.

01

Part 1: Understand AI

What it is, how it works and how to talk to it. The foundation anyone needs, without jargon.

For everyone
  1. 1What AI really is (and what it isn’t)What sits behind Claude, ChatGPT or Gemini: how a model is built, why it makes things up, what it is good for today, and the mental model you will use for the rest of the course.
  2. 2Context is everythingWhat the AI actually sees when you write to it, why the same question gets different answers, what it remembers and what it doesn’t, and how to build a context document for your business so you stop explaining it every time.
  3. 3Claude, ChatGPT, Gemini: how to chooseThe landscape as of October 2026: what each assistant offers, what changes between free and paid, which model to pick inside the app, and how to settle on a main tool for you and your team.
  4. 4Prompts that work the first timeRole, context, task, format, constraints, examples and success criteria. The techniques Anthropic recommends, how to fix a draft without starting over, and how to build your prompt library.
  5. 5Privacy, risk and judgmentWhere what you paste into a chat actually goes, a traffic light for what to share, Colombian data law in plain terms, the risks beyond privacy, and who answers for what AI does.

02

Part 2: AI at work

Projects, connectors, automations and how to bring it to your team with a method. No code.

For everyone
  1. 6Projects, files and connectorsStop repeating context: a project with your context document and fixed instructions, which files to upload and what the AI does with them, artifacts, and connectors (MCP) to your email, Drive and calendar, with the permissions set right.
  2. 7Research, writing and analysis with AIConcrete recipes for sourced research, writing that keeps your voice, spreadsheet analysis with real calculations, contract comparison and turning meetings into tasks.
  3. 8Automations: when AI works on its ownScheduled tasks, n8n, Zapier or Make with an AI step, WhatsApp agents and invoice data extraction. What to automate first, how to do it with a person approving, and what not to automate.
  4. 9Trust, but verifyHow to build checks into every workflow: what to check by output type, a test set of real cases, where the approving person goes and which red flags force a review.
  5. 10Bringing AI to your teamHow to go from everyone using ChatGPT their own way to a shared method: a 30, 60 and 90-day plan, short hands-on training, what to actually measure, and how to talk to people about the changes.

03

Part 3: Claude Code

An agent that builds for you. For anyone who wants to create their own tools, developer or not.

Builders track
  1. 11Claude Code: first steps without being a developerWhat a coding agent is, the terminal in ten minutes, installing Claude Code and your first session building something real.
  2. 12CLAUDE.md, memory and context managementHow to teach it your project once, keep sessions clean and stop the agent from getting lost in long tasks.
  3. 13Planning, permissions and working with GitPlan mode, permissions, branches, pull requests and review. The safe workflow for letting an agent change real things.
  4. 14Skills, commands, hooks and subagentsTurn what you repeat into skills and commands, enforce rules with hooks and split work across specialized subagents.
  5. 15MCP, the cloud and agents that work while you sleepConnect Claude to your tools with MCP, run cloud sessions, schedule routines and let an agent look after your pull requests.

04

Part 4: Agentic engineering

Designing, specifying, orchestrating and evaluating agent systems that do real work.

Builders track
  1. 16What an agent really isA model, some tools and a loop. When a fixed workflow fits, when an agent does, and why simple almost always wins.
  2. 17Writing tasks for agentsSpecs, acceptance criteria and tests as the contract. How to write a task an agent can finish and you can verify.
  3. 18Orchestration: many agents at onceParallel agents, workflows and cross-review. When it is worth the cost and when it is just extra complexity.
  4. 19Building with the API and the Agent SDKFrom using tools to building them: the Claude API, custom tools, the Agent SDK and a first agent for your business.
  5. 20Evals, cost control and failure modesHow to know your AI system works: evaluations, monitoring, cost per task and the most common failure modes.
  6. 21Capstone: from idea to productionPick a real process in your business and take it from idea to a working system, using everything from the course.

24 · Free

Official resources from Anthropic

Anthropic’s official courses, organized by level and linked to each part of this course.

See all resources

89 terms

AI glossary

The course’s vocabulary in plain English, with examples. Look a word up whenever one slows you down.

Glossary

Who it is for

  • Owners and managers who want to understand AI well enough to make good decisions, not just follow the hype.
  • Teams already using ChatGPT or Claude each in their own way who want a shared method.
  • Non-technical people who want to build their own tools with Claude Code.
  • Developers who want to go from using an assistant to designing agent systems.