My Boss Now Requires AI at Work — What Should I Do?

Work & career Guide6 min read·Updated July 4, 2026
The short answer

Start with the AI tools your company has approved, never paste confidential client or company data into free public AI tools, and give yourself a few weeks to practice on low-stakes tasks. Most people feel comfortable within a month of regular use.

Getting a directive to "use AI" at work can feel like being handed a tool you've never seen before and told to build something with it — today. The good news: this is much less complicated than it sounds, and most people feel comfortable within a few weeks. Here's how to approach it calmly.

First, find out what tools you're actually supposed to use

"Use AI" means very different things in different workplaces. Before you do anything else, clarify with your manager:

  • Which specific tool? Is the company providing Microsoft Copilot, Google Gemini, a custom internal tool, or asking you to use your own judgment?
  • Is there a license or account provided? Or are you expected to use free tools on your own?
  • Is there written guidance? Many companies have a policy document — ask HR or your IT department if one exists.

Knowing exactly what tool you're expected to use matters a lot, because different tools have very different rules about what happens to your data.

What you should never paste into a free public AI tool

This is the most important safety rule, and it applies even if you're asked to use AI and given no other instructions.

Never paste the following into a free AI tool like the public version of ChatGPT or a free Google account:

  • Client names, addresses, or any identifying information
  • Internal financial data, sales figures, or unreleased product information
  • Employee personal information or HR records
  • Legal documents under confidentiality
  • Passwords, API keys, or any credentials

Free AI tools may use what you type to improve their models. A company-approved tool — like Microsoft Copilot inside your organization's Microsoft 365 tenant, or Google Gemini for Workspace — has a contract that covers this differently. If you're unsure which category your tool falls into, ask your IT department before pasting anything sensitive.

How to get up to speed quickly

You don't need to become an expert. You need to find two or three tasks in your regular job where AI genuinely saves you time, and build from there.

Start with low-stakes writing tasks

AI tools are most reliable at drafting and editing text. Good places to start:

  • Summarizing a long email or document you received
  • Writing a first draft of a routine email or memo
  • Cleaning up meeting notes into bullet points
  • Generating a list of ideas you then review and trim

These tasks are easy to verify — you can read what the AI produced and catch anything that's wrong before it leaves your desk.

Use what's already built into your tools

If your company uses Microsoft 365, Copilot is probably already in your Word, Excel, and Outlook. If you use Google Workspace, Gemini is built into Docs and Gmail. These are the safest starting points because they're already approved and connected to your existing files.

Ask to watch someone else first

If a colleague already uses AI regularly, ask if you can watch them work through a task for ten minutes. Seeing it in action answers more questions than reading about it.

How to talk to your manager about concerns

If you're worried about data privacy, accuracy, or being asked to do something that feels off, it's worth raising it directly.

A simple way to frame it: "I want to make sure I'm using this the right way — can you point me to our data policy for AI tools, or let me know what's off-limits to paste in?" Most managers will appreciate the question. It shows you're taking it seriously.

If you're asked to use AI to generate content that will be presented as original expert work without any human review, that's worth flagging. In most professional contexts, you are still responsible for what you submit, regardless of how it was drafted.

What to try next

Once you know your tool and your guardrails, the next step is learning to give it better instructions. Most of the frustration people feel with AI comes from prompts that are too vague. The guide on AI prompts for office work has copy-paste examples you can use for common work tasks. If your company uses Microsoft tools, Copilot in Word and Excel walks you through the specific buttons and panels step by step.

Published July 4, 2026 · Updated July 4, 2026How we test →

Frequently asked questions

What if I'm worried about AI replacing my job?
That's a real concern worth taking seriously. Right now, AI is much better at specific tasks — drafting, summarizing, formatting — than at the full context and judgment your job probably requires. Learning to use it well makes you more valuable, not less.
Can I refuse to use AI at work?
That depends on your employment contract and local labor laws. In most cases, if a company-approved tool is provided and training is offered, refusing isn't straightforward. If you have a specific objection — a disability accommodation, a religious concern, or an ethical issue — talk to HR directly.
What's the difference between ChatGPT and a company-approved AI tool?
A company-approved tool has a contract with your employer that covers data privacy, meaning your content is less likely to be used to train public models. Free tools like the public version of ChatGPT have different terms — what you type there may be used to improve the AI.
How long does it take to get comfortable with AI at work?
Most people feel noticeably more confident after two to four weeks of using it on real tasks. The learning curve is much shorter than most people expect.
What if the AI gives me wrong information?
It will, eventually. AI tools can sound confident while being wrong, especially on specific facts, numbers, or recent events. Always verify anything important against a reliable source before using it in work you submit.
Radim Sekera
Founder & editor

Radim is a software developer who spends his days building with AI and his evenings explaining it to family members who don’t care how it works — only what it can do for them. Every guide is tested by hand before it’s published.