Before You Tap That App
Last time, we looked at what AI can and can't do on the shop floor. So you might be itching to open a chatbot and get going. Hold on a second, though.
Here's a question for you. If a customer handed you a tenner, would you put it in the till or in your own pocket? The till, obviously. It's the same money either way. What changes is who's responsible for it and what rules apply.
AI tools work in a similar way. The same brand, whether that's ChatGPT, Gemini, Claude or Copilot, can come as a free personal account or as a work account your employer manages. They might look almost the same on screen, but what happens to what you type can be quite different.
In this level, we'll find your workplace rules, compare work and personal accounts, and build a quick check you can run before you open any tool.
Step One: Find Your Workplace Policy
So where do you look? Start with the places your employer already keeps its rules. Try the staff handbook, the intranet or staff app, and any IT "acceptable use" or data protection policy. Some retailers now have a separate AI policy, and others have added a section to an existing one.
Can't find anything? That's common, and it's fine. Ask your line manager, your IT helpdesk or whoever looks after data protection. If you can, ask in writing. A short message works well: "Is there an approved AI tool I can use for writing product descriptions, and are there any rules I should follow?"
Until you get an answer, play it safe and don't put work information into any AI tool.
Once you've found the policy, summarise it in a few lines. You're looking for answers to these questions:
- Which tools are approved?
- Which account should you use: work, personal, or either?
- Which tasks is AI allowed for?
- What information is off limits? (We'll go into this properly next level.)
- Should you say when you've used AI?
- Who do you ask when you're unsure?
If you can answer those six, you understand your policy better than most people do.
Work Account vs Free Personal Account
This is where the till-or-pocket idea really matters. Let's look at what the providers say on their own help and privacy pages.
ChatGPT. On a personal account, a setting called "Improve the model for everyone" controls whether your conversations can be used to train OpenAI's models. You can switch it off under Data controls. For ChatGPT Business and ChatGPT Enterprise, OpenAI says it doesn't train its models on your business data by default.
Claude. On Free, Pro and Max accounts, users choose whether their chats can be used to train Anthropic's models. Anthropic says those consumer settings don't apply to services under its commercial terms, such as Claude for Work.
Gemini. Google's privacy page for the consumer Gemini app asks people not to enter confidential information they wouldn't want a reviewer to see. With a qualifying Google Workspace edition, Google says your content isn't reviewed by humans or used to train models outside your organisation without permission.
Microsoft 365 Copilot. Under Microsoft's "enterprise data protection", prompts and responses aren't used to train its foundation models. Copilot also follows your organisation's permissions, retention rules and admin settings.
Can you see the pattern? The difference isn't only about training. It's about control. With a work account, your employer has signed a business agreement, and their admins manage the settings, how long data is kept and who can see what. With a personal account, you agreed to the terms. Your employer can't see what you shared, can't delete it and can't protect it.
That's why "it's the same app" isn't a good enough reason. The account you're signed into matters.
One more thing. These terms do change. Before you rely on any of this, check the provider's current help page or ask your IT team.
The Three-Question Check
Here's a simple habit. Before you open a tool for a work task, ask yourself three questions in this order:
- Is this tool approved? If it's not on the list, stop there.
- Am I signed in with my work account? It's an approved tool only when you're using the approved version.
- Does the policy allow this task? Some employers allow AI for writing signs but not for anything involving customers or staff.
If you answer "no" or "not sure" to any of them, stop and ask.
Let's try it out. Priya, an e-commerce assistant, wants to rewrite a product description. Her company provides Copilot through work, she's signed in with her work email, and the policy says AI can be used for drafting listings. Yes, yes, yes. Off she goes.
Now Sam, a supervisor, wants help with next week's rota. He opens a free chatbot on his own phone and starts typing in staff names and availability. That's a no at question one, and again at question two. Even if the result looks great, Sam has moved staff information somewhere his employer doesn't control.
What should Sam do instead? He should check whether there's an approved tool for planning, or ask his manager.
Try It on Your Own Work
This week, find your employer's policy or send that short question to your manager. Then write your six-line summary and keep it somewhere handy, like a note on your phone.
Checklist:
- I know where our AI or IT policy lives, or I've asked who to check with
- I know which tools are approved
- I can tell whether I'm signed in to a work account or a personal one
- I run the three-question check before each work task
- I know who to ask when I'm unsure
Key Takeaways
- Find your workplace policy first. If there isn't one, ask in writing and hold off until you hear back.
- The same AI brand can come as a personal account or a work account, and they handle data differently.
- Work accounts put your employer in control of settings, retention and access. With a personal account, the agreement is between you and the provider.
- Provider terms change, so check current help pages rather than relying on memory.
- Before every task, ask: Is it approved? Am I on my work account? Does the policy allow this task?