You’ve got a good personal AI account, and lately you’ve been using it for real work — pasting a customer’s contract in for a plain-English summary, dropping in employee numbers to sort a schedule, asking for help with a touchy email about someone’s performance. Good. That’s exactly the kind of use that saves you hours. It’s also where people get themselves in trouble.
The risk isn’t using AI at work. It’s pasting the wrong things into the wrong account. The good news: staying safe is a couple of simple habits, not swearing off the tool.
The one gut-check
Before you paste anything sensitive, ask yourself: “Would I be fine if this showed up company-wide — or on a billboard — with my name on it?” If the answer’s no, don’t paste it as-is. That single question catches almost everything that matters: client data, employee records, financials, anything under an NDA.
Do this
Strip the identifiers. Most of the time you need the AI’s help, not the real names. Swap “Dave Johnson, account #4471, $18,200 past due” for “a customer who’s $18k past due.” You get the same useful draft without handing over the sensitive part.
Turn off training/history. Most tools have a setting that stops your conversations from being used to improve the model. Switch it on for any account you use for work — two minutes, once.
Know the bright lines. Customer personal info, employee records, anything NDA-covered, financials that aren’t yours to share — those don’t belong in a personal free account, period. Use a company-approved tool, or strip them first.
Why it matters
This isn’t hypothetical. People have been disciplined and fired for pasting confidential data into a personal AI account — sometimes without realizing the tool held onto it. You don’t have to be paranoid; you just have to not be careless. The habits above take the risk from “career problem” down to “non-issue.”
If you’re the one who signs the checks
This is where a simple, written AI Use Policy earns its keep. Your team is going to use AI — whether you’ve blessed it or not — so a one-page policy that spells out what’s okay, what’s off-limits, and which tool to use turns a liability into a system. (There’s a fill-in template in the Context Starter Pack to get you going.)
This is what lets you scale it
Here’s the upside most people miss: getting this right isn’t about restriction, it’s about permission. Once you’ve got the habits and a clear policy, you can confidently put AI in front of your whole crew without the horror stories — which is when it stops being your personal trick and becomes a company-wide advantage. The boring safety step is what unlocks the big one.
Carry this out
Use AI at work — freely. Just run the billboard test before you paste, strip what’s sensitive, and keep the truly confidential stuff in an approved tool.
Reply and tell me: what’s the one thing you weren’t sure was safe to put into AI? I’ll give you a straight answer.
New here? The fastest way to put all of this to work is the free Context Starter Pack — one folder, three files, one hour.