Start Here: You Already Use AI
This content was developed with AI assistance and is regularly reviewed for accuracy.
Snapshot: October 2026. Model names, prices, and product features change monthly; What's New in AI is the maintained chronology.
You do not need an introduction to artificial intelligence. If you have asked ChatGPT, Claude, Gemini, or Copilot a question this week, you are already a user. What most people have not learned is how to get reliable work out of these assistants: what to hand off, how to describe the job, what to give the assistant so it does not guess, how to catch its mistakes, and how to keep it from doing something you cannot undo.
That gap is what this track closes. AI Fluency is not a course about what AI is. It is a short, practical path to using the assistant you already have the way a capable colleague would: as something you delegate to, brief, equip, and check.
Learning focus
By the end of this track you will be able to:
- Decide which tasks to delegate to an assistant, which tool and mode to use, and when to keep the work human
- Write a brief that an assistant or agent can act on without guessing
- Supply context through files, projects, memory, connectors, and skills instead of re-typing it
- Verify output, recover when a task goes wrong, and keep agents away from irreversible actions
Who this is for
This track is written for people who use AI at work or at home and want better results, not for engineers building AI products. You do not need programming knowledge or math. You need an account on at least one assistant and a willingness to try each technique on a task that matters to you.
If you have never opened an AI assistant, start anyway. The first lesson explains how the current tools behave, and the Hands-On AI Exercises page has a 15-minute exercise that works on any free account.
If you already run agents, write skills, or build with APIs, you are past this track. Go to the AI Agents Overview instead.
A two-minute readiness check
Answer these four questions honestly. They decide where you should start, not whether you are "ready"; everyone is ready.
Have you used an AI assistant for a real task, not just a test question? If yes, skip nothing but expect the first lesson to move quickly. If no, do the first exercise on the Hands-On page before reading further, so the lessons have something to attach to.
Do you know which account tier you are on? Free tiers now run the smaller model of each family and, on ChatGPT, show ads. That is fine for learning, but it changes what "the AI got it wrong" means. The Decide lesson covers this.
Have you ever pasted a long document into a chat and watched the answer get vaguer? If yes, the Supply Context lesson will be the most valuable twenty minutes in this track.
Has an assistant ever done something you did not expect, such as rewriting more than you asked or sending a message? If yes, read Protect before you connect any account to an agent. If no, read it anyway; the incidents it describes happened to professionals.
The arc of the track
The track follows the order in which a good delegation actually happens. Each lesson takes 15-25 minutes to read and about as long to practice. The whole arc is two to three hours, not weeks.
- How AI Assistants Work in 2026 gives you a working model of today's tools: tokens and million-token context, always-on reasoning and the effort setting, web search and months-old cutoffs, memory, and where errors cluster.
- Decide is the step most people skip: should AI do this at all, which tool, and which mode (fast, thinking, research, or agent).
- Brief replaces the 2023 prompt template with a six-field brief: objective, context, resources, constraints, output, verification.
- Supply Context covers the biggest lever beginners leave unused: attached files, projects, memory, connectors, and reusable skills.
- Check teaches verification habits, when to restart instead of correcting, and how to review what an agent did, not just what it said.
- Protect covers privacy, permissions, prompt injection, and irreversible actions.
After the arc come three short applied pages on making things with AI: writing and documents, images, video, and audio, and an app without code. Then the Hands-On AI Exercises turn each lesson into practice, and Projects asks you to build one reusable thing with an agent.
What you will be able to do
The test of fluency is not knowing vocabulary. It is being able to sit down with a task you have never automated and finish it well with an assistant. Concretely, after this track you should be able to:
- Turn "help me with this report" into a brief that names the source files, the audience, the length, and how you will check the result
- Run a research task in a research or agent mode, then spot-check two citations before trusting any of them
- Set up a project or custom instructions so you stop re-explaining your job every session
- Connect one account, such as a calendar or mail, with the narrowest permission that still does the task
- Recognize the two failure patterns that caused the worst 2026 incidents: an agent with too many keys, and an agent reading text that tells it what to do
Notice what is not on that list: memorizing which model is best this month. Those facts change monthly and live in What's New in AI and the AI Resources Hub. This track teaches the habits that survive the next model release.
How to get the most from it
Work on a real task as you read. Pick one thing you do every week, such as a status update, a lesson plan, a client email, or a budget check, and use it as your running example through all six lessons. Every exercise on the Hands-On page accepts your own task in place of the sample.
Use whichever assistant you already have. The techniques are the same across ChatGPT, Claude, Gemini, and Copilot; where a feature has a different name in each app, the lessons say so. A paid tier helps for the research and agent exercises, but nothing in the core arc requires one.
Expect the tools to change under you. Every page in this track carries a snapshot date, and anything that names a product or price is written so you can tell what was true when, and where to check what is true now.
Key takeaways
- You already use AI; this track teaches delegation habits, not definitions
- The arc is decide, brief, supply context, check, protect, in that order, in two to three hours
- Work on one real task of your own throughout; it makes every lesson stick
- Model names and prices are deliberately kept out of these lessons and live in What's New in AI
Next steps
Continue to: How AI Assistants Work in 2026
- Want to try something before reading? Start with Hands-On AI Exercises
- Prefer a guided walkthrough of one real task? Open the playbook Your First Delegated Task