Skip to content
ElementaryMBA
Browse Catalog
Live
Instructors
Sign in
☰
←
AI Operations & Automation for Small Business
Lesson 12 of 12
Lessons
1 · Spotting the tasks actually worth automating
2 · Mapping a simple workflow before you automate it
3 · No-code automation and AI assistants for everyday ops
4 · AI customer support — and when a human must step in
5 · AI help with inventory, data, and admin
6 · Practice: spot it, map it, build it
7 · The real cost of tools — subscriptions, tokens, and time saved
8 · Reliability and a fallback for when AI is wrong
9 · Handling customer data responsibly
10 · Keeping a human accountable
11 · Capstone: your automation plan
▸
12 · Check your understanding
12 · Check your understanding
↗
Share
1. Which task is the BEST candidate to automate in a small business?
Deciding whether to give an angry customer a refund
Sending the same order-confirmation reminder you send for every order
Signing a legal contract
Handling a sensitive customer complaint
2. Why should you map a workflow before buying an automation tool?
Because tools are cheaper on weekends
Because you can't automate a process you can't describe — and automating a messy process just gives you a faster mess
Because mapping replaces the need for any tool
You shouldn't; buy the tool first
3. In an automated workflow, what's the division of labor between a no-code tool and an AI assistant?
They do the exact same thing
The no-code tool handles the 'when X, do Y' plumbing; the AI assistant handles language (writing, summarizing)
The AI assistant does the plumbing; the tool writes the emails
Neither needs a human to approve anything
4. When should an AI support bot automatically hand off to a human?
Never — that's the point of a bot
Only after midnight
When the customer is upset, asks twice, asks for a person, or the issue involves money or something sensitive
Only for questions about store hours
5. AI cleaned up your inventory spreadsheet in seconds. What should you do before relying on it?
Nothing — AI doesn't make mistakes on data
Spot-check the result, because a wrong count means a missed sale or a stockout and a wrong number in your books is a real problem
Delete your original data immediately
Publish it as your official records unread
6. When an automation calls an AI model's API, what do you actually pay for?
A one-time flat fee forever
Per token — both your input and the model's output — so a busy automation can run up a real bill
Per customer only
Nothing; AI APIs are free
7. What does it mean to design an automation for the fact that AI will sometimes be wrong?
Assume AI is always right and skip checks
Keep AI off the high-stakes path alone, route low-confidence cases to a human, test on real cases, and build a working fallback
Turn off all monitoring once it works once
Let the bot decide refunds without review
8. Which is a responsible way to handle customer data, per the FTC?
Collect everything 'just in case' and keep it forever
Take stock of what you collect, keep only what you need, lock it down, delete what you don't need, and don't paste it into public AI tools
Paste customer details into any free AI tool to save time
Assume your automation tools handle security so you don't have to think about it
9. Your automation sent a customer a wrong charge. Who is accountable?
The tool — 'the AI did it' settles it
Nobody, because it was automated
You — a named human owns the automation, fixes the mistake, and makes it right; 'the AI did it' is not a defense
The customer, for trusting a bot
10. What's the honest one-line goal of automating your operations?
Remove every human from the business
Automate the boring, repetitive work so you have more time for what matters — while a human maps, checks, and stays accountable for it
Use as many AI tools as possible
Hide that you use AI from customers
Submit answers
Sign in to track your progress
← Previous
12 lessons to finish
🐞 Report a problem
12 · Check your understanding · ElementaryMBA