For Security Guards ·
What you'll accomplish
By the end of this guide, you'll have a way to practice handling your toughest recurring security situations — before they happen. Claude acts as the difficult person, you practice your response, and you get immediate feedback on what worked and what to do differently.
What you'll need
Go to claude.ai on your phone or browser. Log in and start a new chat by tapping the "New conversation" button (pencil icon or "+" sign). You're going to set up a role-play scenario.
What you should see: A blank chat interface ready for your message.
In your first message, tell Claude exactly what role-play you want and what role it should play. Be specific about the scenario — the more detail you give, the more realistic the practice.
Type something like this:
"I want to practice de-escalation as a security guard. You will play the role of an angry tenant who has been told they cannot enter the building because their building access was suspended for non-payment. They are frustrated, becoming louder, and are making a scene in the lobby. Be realistic — push back when I give weak or escalating responses. When I give a good de-escalation response, acknowledge it. Start the scenario now."
What you should see: Claude starts the scenario in character — playing the angry tenant.
Reply to Claude as if the situation were real. Don't break character — respond with what you would actually say. For example:
You: "Sir, I understand you're frustrated. My name is Officer Johnson. Can we step to the side and I'll explain what's going on?"
Claude (in character) will respond — maybe the person calms down a bit, maybe they push back harder. You continue the conversation.
After you resolve (or fail to resolve) the scenario, ask Claude: "Step out of character and tell me what I did well and what I could have done differently."
Claude will give you specific feedback — what phrases escalated the situation, what de-escalation language worked, and what you could try next time.
What you should see: Honest, specific feedback on your language and approach.
Reset the scenario and try the same situation with different language — test a different opening, a different tone, or a different offer (like calling a supervisor to help). Practice until you have 2–3 reliable responses you feel confident using.
For angry tenant scenarios:
Play an angry tenant who was given a parking ticket by building management and is demanding I remove the ticket or "fix it." I'll practice staying calm and firm while explaining I have no authority over parking enforcement.
For trespassing scenarios:
Play a homeless individual who returns to this property every day. They know it's private property but argue they're not bothering anyone. Practice with me until I find language that gets them to leave voluntarily without it escalating.
For medical situations:
Play a visitor who is becoming increasingly agitated in the lobby, possibly in a mental health crisis. They're not making threats but their behavior is erratic. I'll practice a calm, supportive approach to get them seated and get help.
For feedback after any scenario:
Step out of character. What specific phrases did I use that escalated or de-escalated the situation? What would you recommend I say differently?