Skip to content

Lesson 003Field guide · AI Basics

Types of AI Tools

Chat assistants, image tools, copilots — what each kind is actually for.

Beginner10–13 min5 sections · exercise · quick check

You'll learn

  • The three categories that cover almost every AI tool
  • Why chat assistants are the right place to start
  • What built-in AI copilots do well — and where they're narrow
  • What generator tools are for, and why their results vary more
  • The pick-one rule that beats tool-hopping

01Section

One skill, many wrappers

New AI tools launch every week, each announced like it changes everything. Trying to keep up with the names is a losing game — and the good news is you don't have to play it.

Almost every AI tool you'll meet is one of three kinds: a chat assistant, an AI feature inside software you already use, or a generator that makes images, audio, or video. Behind the different logos sits the same kind of pattern engine you met in Intro to AI. The interface changes; the skill of using it doesn't.

A useful analogy

Think of cars. Hundreds of models exist, but once you can drive one, you can drive nearly all of them — and knowing the categories (sedan, truck, van) tells you what a vehicle is for faster than knowing its brand name ever could.

Key takeaway

Learn the categories, not the product names. Names change every month — the categories, and your skills, carry over.

Checkpoint

02Section

Chat assistants: the general-purpose tools

ChatGPT, Claude, Gemini, Copilot — these are chat assistants: open a conversation, type what you need, get a response. They're the general-purpose category, built for thinking work of every kind: drafting, summarizing, explaining, planning, brainstorming, comparing options.

This is where a beginner should start, for a simple reason: everything you've practiced in this path so far — and everything in the rest of it — works in any one of them. One tool, one skill set, nearly unlimited range of tasks.

Where to start

If you only ever learn one AI tool, make it a chat assistant. Every other category is a specialist; this one is the generalist.

Checkpoint

03Section

AI inside tools you already use

The second category doesn't live in its own app. It's the AI features appearing inside software you already open every day:

  • Your email app suggesting replies or drafting messages
  • Your word processor summarizing or rewriting a document
  • Your spreadsheet app helping build a formula
  • Your meeting app producing notes and action items
  • Your design app removing backgrounds or extending images

Plain-language definition

Copilot the common name for an AI assistant built into another piece of software, there to help with that software's specific job. (Confusingly, Microsoft also named its own assistant Copilot — the lowercase idea is bigger than any one product.)

The trade-off is straightforward. Built-in AI needs no setup and sits exactly where the work happens — but each one does a narrow job. Your email copilot won't plan your week, and your spreadsheet copilot won't draft a difficult message. A chat assistant can do a rougher version of nearly all of it.

Quick knowledge check: Compared with a chat assistant, what's the main trade-off of an AI feature built into an app you already use?

Quick knowledge check

Compared with a chat assistant, what's the main trade-off of an AI feature built into an app you already use?

Checkpoint

04Section

Generators: image, audio, and video tools

The third category makes media instead of text: describe what you want, and the tool generates an image, a voice recording, music, or a video clip. People use them for illustrations, mockups, social graphics, voiceover drafts, and short video.

Expect a bumpier ride than with chat. When a text draft is 80% right, you edit the last 20% in seconds. When an image is 80% right — wrong hands, odd lettering, almost-your-logo — you usually can't fix it directly; you re-describe and regenerate. Words are an imprecise way to specify a picture, so plan on more attempts to get what you pictured.

What a generator prompt looks like

"A warm, simple illustration of a corner bakery at sunrise, flat colors, no text in the image." Same idea as a text prompt — describe the subject, the style, and what to avoid — but the result takes more rounds to land.

One thing to know before using generated media

The rules on using generated images, audio, and video — commercially or otherwise — vary by tool and keep changing. Before generated media represents you or your business, check the tool's current policy.

Checkpoint

05Section

Pick one, learn it well

Here's the rule that protects you from the weekly wave of new launches: one chat assistant learned deeply beats five tools sampled shallowly. Skill compounds inside a tool you know — tool-hopping resets you to beginner every time.

  1. Pick one chat assistant: Any of the major ones. Which you pick matters far less than sticking with it.

  2. Use it on real tasks for a few weeks: Not demos — your actual emails, notes, and plans. That's where skill forms.

  3. Add a tool only when a task demands it: Need an actual image? Now try a generator. Live in spreadsheets all day? Try that copilot. Let the task pull the tool in — never the hype.

Key takeaway

Depth first, breadth only when a task demands it. "Should I try this new tool?" has a default answer: not until a real task needs it.

Pause and think: Think of the task you most want AI help with this week. Which of the three categories does it actually call for?

Checkpoint

Prompt exercise

Run the same prompt in two tools

Copy this prompt into ChatGPT, Claude, Gemini, Copilot, or whichever AI tool you have access to — then run the exact same prompt in a second chat assistant (most have free versions). If you only have access to one, run it twice in two fresh chats. The website doesn't run AI itself — this is a follow-along exercise in your own tools.

Act as a practical AI coach for a beginner. Suggest a simple 5-day starter plan for getting comfortable with a chat assistant: one small, specific task to try each day, each under 15 minutes, using things I already do (emails, notes, plans, questions). For each day, say what the task teaches me. Plain language, no jargon.

Reflection: Compare the two answers. Different wording, probably — but likely the same level of usefulness. That's the lesson: your prompting skill transfers across tools, so pick one and go deep.

Quick check

3 quick questions — no pressure

There's no pass or fail here. Answer them all, and we'll show you the answers either way.

1. A colleague wants AI to turn rough meeting notes into a clean summary. Which category fits that job?
2. What's the main reason results from image generators vary more than results from chat assistants?
3. A new AI tool is all over your feed this week. According to this lesson's rule, when should you try it?