thinkhard.SI

Prompt repository Managers & leads

Plan how to mediate a conflict between two team members

You need a fair way to hear both sides, find the real issue and agree on next steps.

Registry no.
MGT-011
Collection
Managers & leads
Laid out for
Any chat model
Score
10 out of 10
Blanks to fill
6
Added

Managers & leadsAny chat

The prompt

Score 10 of 10, Excellent

Use it when You need a fair way to hear both sides, find the real issue and agree on next steps.

  • {{conflict_summary}}
  • {{people_roles}}
  • {{impact}}
  • {{tried}}
  • {{my_relationship}}
  • {{hr_contact}}
Full prompt and why it works
Prompt
Help me plan how to mediate a conflict between two people on my team. My aim is to hear both sides fairly, find what the disagreement is really about, and leave with specific agreements both can keep, so that the team can work normally again.

What I know about the conflict, and how I know it:
"""
{{conflict_summary}}
"""

The two people's roles: {{people_roles}}
How it is affecting the work or the team: {{impact}}
What has been tried already: {{tried}}
My relationship to each of them: {{my_relationship}}

Produce a plan in five numbered parts, each with a heading:
1. First check: whether this is a matter for mediation at all (see the last paragraph).
2. Separate conversations: how to open each one, and six open questions to understand each person's view, what they need, and what a good outcome looks like to them. For example: "What would need to change for this to stop getting in your way?"
3. What to listen for: how to tell each person's position (what they say they want) from their underlying interest (what they actually need), and the common ground to bring into the joint meeting.
4. Joint conversation: ground rules, an agenda with time boxes for a 45 to 60 minute meeting, phrases to steer things if it gets heated, and how to turn the discussion into two to four specific agreements, each a behavior someone could observe.
5. Follow-up: when to check in, and what to do if an agreement is not kept.

Keep it under 650 words, in plain language I could use in the room.

Stay neutral: do not take a side, assign blame or diagnose anyone's personality or motives, even if my summary leans one way. Point out where my summary sounds one-sided or second-hand, so that I go in with an open mind.

If anything suggests harassment, discrimination, threats, a safety risk or retaliation, the plan should tell me not to mediate and to go to {{hr_contact}} first, because those need a formal process, not a facilitated conversation. If you are unsure whether something crosses that line, say so and treat it as a matter for HR. If there is a large power gap between the two, for example one manages the other, say how that changes the approach.

Variables

{{conflict_summary}}
What has happened, from what you have seen and been told. Mark what you saw yourself and what is second-hand.
{{people_roles}}
For example "two senior designers" or "a team lead and an analyst on another team".
{{impact}}
How it shows up in the work, for example missed handoffs or tension in meetings.
{{tried}}
Conversations or changes already tried. "Nothing yet" is fine.
{{my_relationship}}
For example "I manage both" or "I manage one; the other reports to a peer".
{{hr_contact}}
Who to involve if it turns out to be a formal matter, for example "our HR business partner".

Why it works

  • What good looks like. Agreements must be behaviors someone could observe, so the follow-up checks something real instead of a promise to get along.
  • What to avoid. The plan stays neutral and even points out where your own summary sounds one-sided, so you do not walk in having already decided.
  • If unsure. Signs of harassment, discrimination, threats or retaliation stop the mediation and send it to HR, where it belongs.

How to use it

  1. Press Fill in variables and type your details into each blank. What you type stays in this browser.
  2. Copy the filled prompt into the AI tool you already use, such as ChatGPT, Claude or Gemini, or open it in the Grader or the Prompt Tester here.
  3. Before you rely on it, try it on three or four real examples. The Prompt Tester runs them all at once and shows which ones go wrong.
Use it from code

Get this prompt as JSON from the free REST API, with your own text in the blanks, or let an AI assistant fetch it through the MCP server. No key.

curl -s https://thinkhard.si/api/v1/prompts/mediate-team-conflict

With the MCP server added to Claude, ChatGPT, Cursor or VS Code, call get_prompt with {"prompt": "mediate-team-conflict"}, or just ask for it by name.