Questions Managers Should Ask AI Before a Difficult Employee Conversation

Start by questioning your own version of the story

Every manager eventually has a conversation they would rather not have. An employee's performance has slipped. Someone's behavior is affecting the team. A previously reliable employee has become disengaged. Two people aren't getting along, or an employee expects a promotion the manager doesn't believe they're ready for.

Managers now have something available that they didn't have before: AI can help them prepare before they walk into that conversation. It can suggest questions, help identify assumptions, offer another way of looking at the problem or even let the manager practice how a conversation might go.

I think that's a good use of AI. What I wouldn't do is ask it to write the entire conversation and then walk into the meeting with a script. The purpose is to help the manager think through the situation, not take over the conversation.

AI can help you prepare. The conversation still belongs to you.

Ask what you might be missing

Suppose a manager asks AI, “I have a difficult employee who is always negative. Tell me what to say to her.”

There's already a problem with the question. We've decided the employee is difficult. We've decided she's always negative. Then we've asked AI to tell us what to say based on those conclusions.

A better question would be: “I believe an employee's behavior is affecting the team. What should I understand, and what assumptions should I consider before I talk with the employee?”

That's a very different use of AI. Instead of asking it to confirm what you already believe, you're asking it to help you make sure you've thought the situation through.

This may be one of the best questions a manager can ask before a difficult conversation: What assumptions might I be making?

Let's say someone is consistently missing deadlines. Maybe they're disorganized. That's certainly possible. But priorities may be conflicting, expectations may not be clear, the workload may be unrealistic, the employee may need additional training or another department may be causing delays.

The manager's original conclusion may turn out to be exactly right. The point is to make sure you haven't decided what the problem is before you've actually talked to the employee.

Another useful question is, “What information should I have before I go into this conversation?” Look at what expectations were communicated, what actually happened and when, whether this is an isolated incident or a pattern, whether the problem has been discussed before and what specific examples you can provide.

There's a big difference between telling someone, “You're always late with your work,” and saying, “The last three reports came in after the deadline. Can you help me understand what's getting in the way?”

One is a conclusion. The other gives you somewhere to start a conversation.

Consider what the situation looks like from the other side

You don't have to agree with the employee's perspective to consider that another perspective exists.

AI can help a manager think through what the employee might say. Maybe the employee believes priorities changed. Maybe expectations weren't as clear as the manager thought they were. Perhaps the employee doesn't realize how their behavior is affecting other people. Or maybe the manager's assessment is completely accurate.

The point isn't to make excuses for poor performance or behavior. It's to go into the conversation prepared to listen instead of simply waiting for your turn to speak.

I'd also ask, “Could anything about my management approach be contributing to this?” Were expectations clear? Was feedback given soon enough? Did the employee have the resources they needed? Has the manager avoided dealing with the problem until they became frustrated?

Sometimes the answer will be no. Sometimes the manager has done everything reasonably expected and the employee needs to make a change. But a good manager should at least be willing to ask the question.

Know what you want from the conversation

Before walking into a difficult conversation, decide what you're actually trying to accomplish.

Do you need better performance? A change in behavior? Clearer expectations? More information? A development plan? Do you need the employee to commit to specific actions, or are you still trying to understand what the real problem is?

AI can help a manager sort through those possibilities, especially when emotions are already involved. It can also help you think about how the employee might react. The person may agree with you, be surprised, become defensive, get emotional or tell you something you didn't know.

You can prepare for those possibilities, but don't mistake preparation for prediction. AI doesn't know how this particular employee is going to respond.

Be careful about questions that confirm what you've already decided

Some questions make us think more carefully. Others simply ask AI to support a conclusion we've already reached.

Questions such as “Why is this employee so difficult?” “How do I prove this employee isn't performing?” or “Tell me why this person isn't leadership material” don't really help the manager understand the situation. They start with the answer already built into the question.

Try turning the question around. Instead of asking, “Why isn't this employee ready for promotion?” ask, “What evidence should I consider when deciding whether someone is ready for greater leadership responsibility?”

Instead of, “How do I deal with a negative employee?” try, “What should I understand before addressing behavior that I believe is affecting the team?”

Those questions leave room for information you may not already have.

And if the question is, “Should I terminate this employee?” you've probably reached the point where another AI prompt isn't what you need. It's time to involve HR and, when appropriate, legal counsel.

Protect the employee's information

This needs to be part of every organization's discussion about managers using AI.

Managers routinely deal with sensitive information about performance, compensation, health, leave, disciplinary issues, complaints and investigations. Don't assume that because you can paste information into an AI tool, you should.

Managers need to follow their organization's policies about AI, privacy, HR and information security. They also need to know which AI tools are approved and what information is permitted.

In many cases, AI doesn't need to know who the employee is. You can ask, “What questions could help me understand why a previously reliable employee has begun missing deadlines?” without giving the person's name, department, age, medical information or employment history.

Give the tool enough information to help you think about the management problem without unnecessarily exposing the employee.

Know when to stop and call HR

AI can be useful when you're preparing for normal management conversations. There are other situations where managers need professional guidance.

Issues involving discrimination, harassment, retaliation, accommodations, protected leave, investigations, workplace safety, threats, significant disciplinary action or termination may require HR, legal counsel or other appropriate professionals. Managers should know their organization's requirements and when they're expected to escalate an issue.

AI might help you recognize that a situation is more complicated than you originally thought. That's useful. It shouldn't become a substitute for your organization's policies or qualified HR and legal advice.

Then close the tool and have the conversation

Once you've used AI to prepare, you're eventually going to have to stop prompting and talk to the employee.

You probably don't need a beautifully written five-paragraph script. You need to know what you observed, why it matters, what you need to understand and what needs to happen next. Then you need to listen.

That's where the actual management begins. The employee may look confused. There may be a long pause before an answer. Something you thought was a performance issue may turn out to be a communication problem. You may realize you need to ask another question instead of moving on to the next point you planned to make.

AI can't make those adjustments for you because it isn't sitting across from the employee. You are.

If managers are already using AI to prepare for employee conversations, I don't think the answer is simply telling them not to use it. Teach them how to use it well. Use it to question your thinking and prepare for the conversation. Protect employee information. Know when HR or legal needs to become involved. Then speak in your own words and listen to the person in front of you.

One final question may be worth asking before you close the AI tool:

“What might I be missing, and what should I be prepared to learn from this employee?”

If that question helps a manager walk into the room better prepared to listen and make a thoughtful decision, then AI has done something useful.

Sources

National Institute of Standards and Technology (NIST). Artificial Intelligence Risk Management Framework (AI RMF 1.0). NIST's framework provides guidance for organizations using AI and addresses responsible AI use, privacy, accountability, transparency and management of harmful bias. https://www.nist.gov/itl/ai-risk-management-framework

National Institute of Standards and Technology (NIST). AI Risk Management and Human-AI Interaction. This is particularly relevant to managers because NIST discusses how human assumptions and cognitive biases can influence AI use. It also notes that human and AI roles in decision-making need to be clearly defined. https://airc.nist.gov/airmf-resources/airmf/

U.S. Department of Labor. Artificial Intelligence Literacy Framework. Released in February 2026, the framework addresses the AI knowledge and skills workers and employers need as AI becomes more common in the workplace. https://www.dol.gov/agencies/eta/advisories/ten-07-25?