9 Simple Questions on AI – Answered By Hema Dey for Employees & Mid-Level Managers

This is the question keeping people awake at night.
And I want to answer it honestly. Not to make you feel better. Because you deserve the truth.

The short answer: probably not your whole job. But definitely parts of it.

Here is the more honest answer.
AI will not replace you. But someone who uses AI well might.
That is not a threat. That is the most useful thing I can tell you right now.

The jobs AI is changing first:
Tasks that are repetitive. The same thing done the same way every day.
Tasks that process information. Reading documents, summarizing reports, sorting data.
Tasks that answer the same questions over and over. Standard customer service. Standard FAQs.

The jobs AI is NOT replacing:
The person who builds the relationship with the client.
The person who makes the judgment call when things go wrong.
The person who knows the history, the context, and the people.
The person who leads the team through change.
The person who asks the right question — not just gives the right answer.

Think of it like this.
When calculators arrived — mathematicians did not disappear.
They stopped doing the boring calculation by hand and started solving bigger problems.
When GPS arrived — taxi drivers did not disappear.
The best ones got better at customer service because navigation was handled.

AI is the calculator. You are the mathematician.
The question is not will AI replace me.
The question is: how do I use AI to become the person who solves the bigger problems?

This is the right question.
And the fact that you are asking it means you already have the most important thing.
Awareness.
The people who become irrelevant are the ones who stop paying attention.
You have not done that. So let us talk about what to do next.

The three things that make you irreplaceable in the AI era:

01 — Human judgment.
AI can produce a thousand options. It cannot tell you which one is right for this client, this moment, this relationship.
You can.
The more you develop your judgment — the more valuable you become.
Read more. Ask better questions. Seek feedback. Reflect on decisions.
Judgment is the most AI-proof skill you own.

02 — Relationships.
AI can write an email. It cannot build trust over three years of showing up consistently.
It cannot remember that your client’s daughter just started college.
It cannot make someone feel genuinely heard in a difficult conversation.
Relationships are yours. Protect them. Deepen them. They are your most valuable professional asset.

03 — The ability to direct AI well.
This is the new skill. And it is learnable.
The people who thrive in the AI era are not the ones who understand AI technically.
They are the ones who know how to ask AI the right questions and get useful answers.
Think of it like being a great manager.
A great manager does not do every task themselves.
They give clear direction, review the output, and make it better. That is exactly how you work with AI.

Stay curious. Stay human. Stay in the relationship. And learn to direct the tool. That is relevance in 2026.

Good news.
You do not need a computer science degree. You do not need to write code. You just need to start with the right things.

The three AI skills worth learning first — in order:

Skill 01 — How to ask good questions.
This is called prompting. And it is just writing clear instructions.
Bad prompt: Write me an email.
Good prompt: Write a friendly follow-up email to a client who attended our webinar last week. They are interested in our services but have not responded yet. Keep it short and warm. Do not be pushy.
The difference is clarity. The clearer your instruction — the better AI’s output.
Think of it like briefing a junior colleague. The better your brief — the better their work.

Skill 02 — How to review and improve AI output.
AI gives you a first draft. Not a final answer.
Your job is to read it critically. Does it sound like you? Is it accurate? Is it right for this person and this moment?
The skill is editing with intention — not just fixing typos but making the output genuinely good.
This is a human skill. AI cannot do this part for you.

Skill 03 — How to identify where AI belongs in your work.
Look at your week. Find the tasks that are repetitive and predictable.
Drafting similar emails. Summarizing meetings. Creating first versions of reports. Researching topics.
These are your AI opportunities. Start there. Build your confidence. Then expand.

One hour a week practicing these three skills will put you ahead of most people in your organization within 90 days.

Right here. Right now. With something tiny.
The biggest mistake people make with AI is waiting until they feel ready.
That feeling does not come before the start. It comes after.

Your first week — four small steps:

Day 01  Open ChatGPT or Claude. Type this exactly: Hello. I am new to AI. Can you explain in simple terms what you can help me with at work? Read the response. Ask a follow-up question. That is it. You just started.

Day 02  Find one email you write regularly. A follow-up. A weekly update. A client check-in. Ask AI to write a first version of it. Read what comes back. Notice what is good and what needs fixing.

Day 03  Take the notes from your last meeting. Paste them into ChatGPT. Ask it to summarize the key decisions and action items. See how accurate it is.

Day 04  Ask AI one question you genuinely want answered about your job or your industry. Treat it like a very well-read colleague who has time to explain things clearly.

That is your first week.
Four small things. Probably four hours total.
You will know more about AI after those four days than most people in your organization.

The secret is not a big launch. It is a small start that you actually do.

This happens to almost everyone the first time.
You open it. The blank box stares at you. You type something vague. The answer is okay but not very useful. You close it and think maybe AI is not for me.
It is for you. You just needed better starting questions.

Ten questions you can ask ChatGPT right now — at work:

01  I need to write a professional email declining a meeting request politely. Can you help me draft it?
02  Summarize these meeting notes into three key decisions and five action items. [Paste your notes]
03  I am preparing for a performance review. What questions should I be ready to answer?
04  Write a LinkedIn post announcing our new [product/service/achievement]. Keep it professional and under 150 words.
05  I have a difficult conversation with a colleague tomorrow about their performance. What is the best way to approach it?
06  Explain [industry term you just heard] in plain language like I am hearing it for the first time.
07  I need to present this data to a non-technical audience. Help me simplify the key points.
08  Write a brief agenda for a 30-minute team meeting about [topic].
09  I am writing a report about [topic]. What headings and sections should I include?
10  What are three ways someone in my role as [your job title] could use AI to save time this week?

Save this list. Use one today.
The more specific your question — the more useful the answer. That is the whole secret.

Yes. That is exactly what this document is.
And here is the thing about learning AI in normal language.
You do not need to learn how AI works. You need to learn how to work with AI.
Those are completely different things.

AI in the most normal language possible:

Imagine your smartest friend. The one who has read everything, knows a little about everything, and always has time to help.
They write really well. They research fast. They never get tired of your questions.
But they sometimes get things wrong with confidence — so you always check their work.
And they do not know you personally yet. You have to tell them your context each time.
That is AI. Your very smart, very fast, occasionally wrong helpful friend.

How to work with that friend effectively:

01  Tell them exactly what you need. Not write me an email. Write me a warm professional email to a client who has not responded in two weeks. Keep it under 100 words.
02  Give them context. They do not know your company, your client, your situation. The more you tell them — the more useful their help.
03  Always read what they give you. They are smart but not perfect. Your job is to check it before it goes anywhere.
04  Ask follow up questions. Not happy with the first answer? Say make it shorter or make it warmer or give me three different options. They never get offended.

That is it. That is how you use AI in normal language.
No jargon. No courses. No technical knowledge required.
Just a clear question and a willingness to check the answer.

I know.
And I want to say something before we go any further.
That feeling is completely valid.

You are being asked to learn something new while still doing your full job.
You are reading headlines about AI taking jobs while trying to figure out whether yours is safe.
You are sitting in meetings where people use words you do not fully understand.
And you are doing all of this while managing your actual workload, your actual team, and your actual life.

Of course you feel overwhelmed. That is the right response to an overwhelming situation.

But here is what I know after working with hundreds of people through technology transitions.
The overwhelm is always loudest right before things start to click.
And things click faster than you think — when you take one small step instead of trying to solve the whole thing at once.

When you feel overwhelmed — do just one of these:

01  Close all the AI articles and tabs. Open ChatGPT. Ask it one work question. Just one. See what happens.
02  Write down the three AI-related things that worry you most. Then read them back. Often the act of naming the fear makes it smaller.
03  Find one colleague who seems comfortable with AI. Ask them to show you one thing they use it for. Not a class. Not a course. Just one thing from one person you trust.
04  Give yourself permission to not know everything yet. This technology is months old for most people. You are allowed to be in the learning phase.

You do not have to understand all of AI. You just have to understand enough to take the next step.
And the next step is always smaller than the overwhelm makes it feel.

This is one of the most important questions anyone can ask about AI.
Because it shows you understand the real problem.
AI content without a human touch sounds like AI content. And people can tell.

Why AI sounds robotic — and how to fix it:

The problem:
AI loves certain phrases. Delve into. It is worth noting. In conclusion. Ensure. Utilize.
It loves long sentences that say a lot without really saying anything.
It loves being comprehensive when you needed to be specific.
None of these are how you actually talk. And your reader can feel the difference.

The fix — five things to do to every AI output:

01  Read it out loud. Every sentence. If you would not say it in a real conversation — rewrite it. This one habit catches most of the robotic language immediately.
02  Add one thing only you know. A specific example from your experience. A client story. A real observation from your work. AI cannot generate this. You can. And it changes everything.
03  Delete the filler phrases. It is worth noting that. In order to. It is important to remember. These phrases add length and remove personality. Cut them all.
04  Shorten the sentences. AI writes long. You talk short. Short sentences sound like a human. Long sentences sound like a document.
05  Add your opinion. AI gives you information. You give it a point of view. What do YOU think about this? That one addition makes the whole piece yours.

AI gives you the scaffolding. Your voice is the building. Never publish the scaffolding without the building.

This is the question every mid-level manager is sitting with right now.
You have been told to bring AI into your team. Your team has questions. Some are excited. Some are scared. Some are quietly resistant.
And you are trying to lead all of them through something you are still figuring out yourself.
Here is how to do it without losing anyone along the way.

The five things that make team AI adoption work:

01 — Address the fear first. Before the tool.
Before you show your team any AI tool — have an honest conversation about what they are worried about.
Are they afraid of losing their jobs? Say so out loud. Do not pretend the fear is not there.
Then be honest about what you know and what you do not know yet.
Teams follow leaders who are honest. They resist leaders who pretend everything is fine.

02 — Start with a win, not a training.
Do not begin with a course or a webinar or a policy.
Begin with one small win. Find one task your team hates doing that AI can help with.
Show them the time saved. Let them feel the benefit before you ask for the behavior change.
People adopt tools that make their lives easier. Show them the easier first.

03 — Make it safe to try and fail.
Your team will produce bad AI output at first. That is part of learning.
If they are afraid of being judged for a bad first attempt — they will not try at all.
Celebrate the experiments. Laugh at the bad outputs together. Make learning feel safe.
Psychological safety is the fastest AI adoption tool you have.

04 — Give them one use case at a time.
Do not hand your team a list of fifty ways to use AI.
Give them one. This week we are all trying AI for meeting summaries. Next week we try it for email drafts.
One use case at a time. Master it together. Then move to the next one.

05 — Be the example.
Nothing accelerates team adoption faster than watching their manager use AI in a real meeting.
Show them your AI output. Show them how you edited it. Show them the mistakes it made.
Your willingness to learn in public gives everyone else permission to do the same.