Capstone: Your One-Page AI at Work Checklist

You now know the main risks of AI at work and the habits that prevent them. This last lesson pulls everything into one page you can print, pin, or keep open in a tab. Then you will practice it on four short work scenarios.

What You'll Learn

• A one-page checklist for before, during, and after any AI task
• How to fill in the parts that are specific to your company
• How to apply the checklist to real work situations
• What to do when something goes wrong

The checklist

Before you start

• What data? I know the level: public, internal, confidential, personal, or client.
• Which tool? This tool is approved for that data level.
• Right account? I am signed in to my work account, not a personal one.
• Least data? I removed names, numbers, and details the task does not need.
• Right access? I am not connecting extensions, agents, or apps that are not approved.

While you work

• Share the part, not the whole. Only the text the task needs.
• Describe instead of paste when the real text is not needed.
• No new tools mid-task. If I need a new tool, I ask first.

Before you send

• Accuracy. I checked every number, date, name, and source.
• Confidentiality. Nothing from another client or internal-only data slipped in.
• Copyright. Nothing looks copied. Important parts are in my own words.
• Bias. It is fair to the people it describes.
• Disclosure. I said AI helped where my policy, the client, or the reader would expect it.

If something goes wrong

• I stop using the tool for that task.
• I tell my manager or IT what was shared, where, and when.
• I do not try to hide it or fix it alone.

Make it yours

The checklist is general. Add your company's details so you do not have to look them up each time:

• My approved AI tools: _______
• Data levels each tool can use: _______
• Where our AI policy lives: _______
• Who approves new tools: _______
• Who I tell if something goes wrong: _______
• Client rules I need to remember: _______

If you cannot fill in a line, that is a useful finding. It is the question to ask your manager or IT this week.

Practice: four scenarios

Read each one and decide what you would do. Then compare with the suggested answer.

Scenario 1: The sales forecast

You need to turn next quarter's sales forecast into three slides. Your company has an approved business AI tool. The forecast is confidential.

Suggested answer: Check that your policy allows confidential data in the approved tool. If it does, use your work account and share only the figures the slides need. If it does not, describe the structure and use made-up numbers, then add the real ones yourself. Check every figure on the final slides.

Scenario 2: The handy extension

A coworker recommends a free AI extension that summarizes any web page, including your internal dashboards. It asks for access to all websites.

Suggested answer: Do not install it on your work laptop yet. It can read every page you open, including internal and client systems. Send IT a short approval request with the link, the access it asks for, and the task. Ask if an approved tool already does this.

Scenario 3: The client report

An AI tool drafted most of a monthly report for a client. It reads well and includes a few industry statistics you did not provide.

Suggested answer: Check every statistic against a real source, or remove it. Make sure no details from other clients appear. Check your client contract and policy on AI use. If the client would assume a person wrote it, add a short note saying AI helped draft it and you checked the figures.

Scenario 4: The personal account slip

Last week, in a rush, you pasted part of a customer complaint, including the customer's name and order number, into your personal chatbot account.

Suggested answer: Stop using the personal account for work. Tell your manager or IT what you shared and when. They can judge the risk and decide next steps. Next time, remove names and numbers first and use an approved tool.

Use AI to stress test your own plan

You can ask an approved AI tool to review a plan for an AI task against the checklist. Describe the task in general terms, without pasting real data.

Where to go next

• For personal AI safety at home, like what never to paste and how to manage your own settings: AI Safety & Privacy: Stay Safe Using Everyday AI.
• To get more done with AI once you have safe habits: ChatGPT at Work: 10x Your Productivity.
• If you lead a team and need to set the rules: AI for Working Professionals: The Manager's Playbook.
• For the wider view on fairness and responsibility: AI Ethics & Responsible AI.

Key Takeaways

• Use one checklist before, during, and after every AI task.
• Before: data level, approved tool, work account, least data, no unapproved access.
• Before you send: accuracy, confidentiality, copyright, bias, disclosure.
• If something goes wrong: stop, tell your manager or IT, and do not hide it.
• Fill in your company's details on the checklist. Any blank line is a question to ask this week.