Lesson 3 of 4 · about 6 min
Guardrails for small business owners
When you finish, you can say what never to paste into an AI tool, how to check output, and how to be honest with customers about AI.
Stuck? Ask the AI tutorProtect the people who trust you
Your customers gave you their names, phone numbers, addresses and sometimes payment or health details because they trust you. Free AI tools may store what you type, and some may use it to improve their products, depending on the plan and settings. Check the tool's current privacy page and settings. As a rule, do not paste full names, phone numbers, addresses, card numbers, bank details, passwords or employee records.
You can almost always get the same help with details removed. Instead of "Maria Santos at 14 Palm Street owes 300", write "a regular customer owes a late invoice". Replace names with Customer A. Keep a private note that links the label to the real person. Many countries have privacy laws for customer data, such as the GDPR in Europe. Ask a local adviser what applies to you.
Check before you send
AI makes mistakes that look polished. It can invent a law, a price, a review or a supplier. Make a habit of a three-point check before sending: facts (dates, prices, names), promises (did I agree to this?), and tone (would I say this to the person's face?). If the message affects money, safety or a contract, read it twice and slowly.
Never let AI send messages for you without reading them. Never ask AI to write fake reviews or to pretend to be a customer. That harms trust and may break the rules of the platforms you use or the law where you live.
Fairness and your own rules
AI learned from human writing, so it can repeat unfair habits. When you ask for help screening job applicants or writing a tenant message, it may lean on assumptions about age, gender, accent or background. Use AI to build a fair list of questions and criteria, and make the hiring decision yourself using the same questions for every candidate.
For regulated matters, such as tax filings, employment law, food safety rules and insurance claims, use AI only to prepare questions and plain explanations. The accountant, lawyer, inspector or insurer decides. If you are not sure whether a use of AI is allowed in your industry or by your payment or platform provider, check their current rules first.
Being honest about AI, and when to say no
You do not need to announce AI in every email, but you should never lie about it. If a customer asks, say so plainly: "I use an AI tool to help draft messages, and I check every one myself." If you use an AI chatbot on your page, tell people it is one and give them a way to reach a human.
Say no when a task needs a human: a grieving customer, a safety complaint, a legal threat, a worker's private problem. Pick up the phone. Also say no to shortcuts that would fool people, such as fake testimonials or invented urgency. Trust is the one thing your small business cannot afford to lose.
Practice now
- Write your own rule card with five lines: what I never paste, what I always check, who approves money messages, how I answer if a customer asks about AI, and when I use the phone instead.
- Take one old customer message, remove names and details, and rewrite it using Customer A and Customer B.
Use any AI assistant you have. Do not paste private, client or patient information into it.
Check yourself
1. Which detail is safest to include in a free AI chat?
2. A customer asks if you use AI. What is the right answer?
3. Which message should you handle by phone, not AI?
Ask the AI tutor
Ask anything about this lesson, in any language. The tutor is an AI: it can be wrong, and it never needs your personal details.
AI-drafted. Checked by an automated validator and a separate AI reviewer; a person has not read every lesson. Translations and tutor answers are AI-made. Found a mistake? Tell us. AI can be wrong. General information only, not professional advice. Not affiliated with any AI company named. Current University is not accredited.