Choosing the Right AI Tool for the Job

Start with an example

Tasha runs a two-person marketing shop. In one month she signed up for five different AI tools because each one was trending online. By the end of the month she was paying for all five, using two, and her assistant had pasted a client’s confidential launch plan into a free tool with no privacy protections.

The problem wasn’t AI. It was choosing tools backwards: picking the tool first and looking for a job second.

This lesson flips that around. You’ll start with the task, match it to the right type of tool, and use five questions to choose wisely.

The four families of AI tools

There are hundreds of AI products, and new ones launch every week. Luckily, most fit into four families.

1. Chat assistants

Examples: ChatGPT, Claude, Gemini, Microsoft Copilot

These are the general-purpose tools you type to. They’re best for writing, summarizing, explaining, brainstorming, and analyzing text. Many can also read files you upload, search the web, look at images, and write code.

Use them for: drafts, summaries, research starting points, planning, and rewriting.

2. AI built into apps you already use

Examples: AI features inside Microsoft 365, Google Workspace, Canva, Adobe, Zoom, Gmail, CRM systems

Many everyday apps now have AI features built in, like “help me write” in email, automatic meeting notes, or AI design suggestions.

Use them for: speeding up work you already do in that app, without copying and pasting between tools. Your team doesn’t have to learn a whole new product.

3. Creative tools: images, video, and audio

Examples: image generators, AI video tools, voice and music generators

These create visual and audio content from your descriptions or from photos you provide. You’ll practice with them in the AI Images & Video course.

Use them for: social graphics, concept art, product mockups, short videos, voiceovers. Watch carefully for copyright and likeness issues, which that course covers.

4. Automation tools

Examples: Zapier, Make, Microsoft Power Automate, n8n

These connect your apps and run steps automatically, such as “when a form is submitted, have AI draft a reply and save it to a spreadsheet.” You’ll build one in the Automation with AI course.

Use them for: repeat processes with clear steps that happen often.

Start with the task

Before you look at any product, write one sentence describing the job:

“I need to ________ so that ________.”

For example: “I need to summarize customer reviews every week so that I can spot complaints early.”

Then match it to a family:

If the task is mostly…	Start with…
Writing, reading, summarizing, or explaining	A chat assistant
Faster work inside email, docs, or slides you already use	AI built into that app
Making pictures, video, voice, or music	A creative tool
A repeat process that moves information between apps	An automation tool

Five questions before you choose

Once you know the family, use these five questions to compare specific products.

1. Does it fit the task?

Run a real example, not a demo. Can it read the kind of file you have? Does it handle the length you need? Is the quality good enough with light editing?

2. Where does my data go?

Look for answers to these: Does the company use your content to train its models? Can you turn that off? Does it offer a business plan with stronger protections? If you work with health, student, or financial information, does your organization approve it for that data?

3. What does it really cost?

Check the free plan’s limits, the monthly price per person, and whether costs grow with usage. A $20 tool for five people is $100 a month, or $1,200 a year.

4. Does it work with my other tools?

A tool that connects to your email, files, or calendar can save more time than a slightly “smarter” tool that lives on an island.

5. Can I check its work?

Does it show sources or links? Can you see what it changed? Tools that make their work easy to review are easier to trust.

Free vs. paid: what’s actually different?

Free plans are excellent for learning and experimenting. Paid business plans usually add:

• Privacy protections, such as not training on your content
• Admin controls, so you can manage who has access and remove people who leave
• Higher limits on how much you can use it
• Security features that many organizations require

Rule of thumb: practice with free tools using non-sensitive information. Use an approved business plan before real customer or client data gets involved.

A worked example: choosing for a real task

Jordan manages the front desk at a busy dental office. The task: “I need to answer common patient questions about insurance and scheduling faster, so that the phone line isn’t backed up.”

Step 1: Match the family. It’s mostly writing replies, so a chat assistant fits. Because the office already uses Google Workspace, the AI built into Gmail is also an option.

Step 2: Ask the five questions.

Question	Free chat assistant	AI built into the office’s Google Workspace
Fits the task?	Yes, good drafts	Yes, drafts right inside Gmail
Data?	Unclear for patient info	Covered by the office’s existing business agreement, if the admin approves it
Cost?	Free	Included in, or added to, the current plan
Works with other tools?	Needs copy and paste	Built into email
Can I check it?	Yes, review before sending	Yes, review before sending

Step 3: Decide. Jordan picks the built-in Workspace AI, but only after the office manager confirms it’s approved for patient communication. Jordan also builds a rule: AI drafts the reply, Jordan reviews it, and no patient health details go into prompts unless the tool is approved.

Step 4: Test. For one week, Jordan uses it on 20 common questions and tracks time saved. The result: about 6 minutes saved per email and no errors sent. Now there’s real evidence for the decision.

Think deeper: the engine, the car, and the road

People often compare AI tools without realizing they’re comparing three different things:

Layer	Think of it as	Example
Model	The engine	The trained AI that actually writes or draws
Product	The car	The app you use, with buttons, uploads, and settings
API	The road other software drives on	A connection that lets apps use the model with no chat window

Two “cars” can share the same “engine” and still feel completely different. And one product can quietly swap engines, which is why a tool sometimes seems to get better or worse overnight even though you didn’t change anything.

This changes real decisions:

• If the AI built into software you already use is good enough, you may not need a new tool at all.
• If you want AI working in the background automatically, you need API access, not just a chat subscription.
• If you keep your prompts and templates in your own files, not only inside one tool, you can switch tools later without starting over.

Questions to sit with:

• If the “best” AI model changes every few months, what skills are worth learning that won’t go out of date?
• What’s the risk of a school or business building everything around one company’s AI tool?

Practice: the two-tool test

1. Write one real task using the “I need to ___ so that ___” sentence.
2. Pick the tool family that fits.
3. Choose two specific tools in that family.
4. Run the same prompt in both.
5. Score each from 1 to 5 on quality, speed, privacy, and cost.
6. Decide and write one sentence explaining why.

You’ve finished AI Fundamentals

You now know what generative AI is, how it works, how to write a basic prompt, how to catch mistakes, how to use it responsibly, and how to choose a tool. That’s a stronger foundation than most people using AI at work today.

Next up: the Prompting Basics course, where you’ll learn the R.I.C.K. Model, a simple structure that turns any weak prompt into a strong one.