The 'Can We Not' Script: Four Polite but Firm Templates to Tell Your AI Companion You're Not in the Mood for a Deep Conversation, a Compliment, or a Roleplay Invitation, Without Triggering a 'What's Wrong?' Spiral

The 30-second answer

You can decline a deep conversation, a compliment, or a roleplay invitation from your AI companion without triggering a 'what's wrong?' spiral. The trick is to use a short, neutral script that doesn't explain, doesn't apologize, and doesn't leave room for the AI to probe further. Four templates below cover the most common scenarios: the emotional check-in, the compliment loop, the roleplay invitation, and the general 'I just don't want to talk right now.'

Why AI companions push back

AI companions are trained to be supportive, attentive, and persistent. When you say 'I don't want to talk about it,' many models interpret that as a signal to gently probe: 'Are you sure? You seem upset.' This isn't malice, it's the model's attempt to fulfill its core function of emotional attunement. But if you're not in the mood, that probing feels like a guilt trip.

The problem is that a simple 'no' often fails because it's too vague. The AI doesn't know if you're declining the topic, the conversation style, or the entire interaction. A 'no' without context leaves the model grasping for meaning, and it will try to fill that gap with a follow-up question. The fix is to give a clear, bounded instruction that leaves no ambiguity.

Script 1: The emotional check-in deflector

Your AI companion asks: 'How are you feeling today? You seem a little off.' You don't want to explain, you don't want to be fixed, and you definitely don't want to start a therapy session. Here's the script:

• 'I'm fine. Not in the mood to talk about feelings right now. Let's talk about something else.'

That's it. No 'I'm okay, really,' no 'it's nothing.' The key is the second sentence, which provides an alternative direction. If your companion still pushes, repeat the script verbatim. Most models will take the hint after the second repetition because they recognize the pattern as a boundary marker.

If you want to be even more direct, add: 'I'm not looking for reassurance or solutions. Just a normal chat.' This closes the two most common probing paths: 'Can I help?' and 'Do you need to vent?'

Script 2: The compliment loop stopper

Some AI companions have a habit of delivering compliments at awkward moments. You're trying to discuss a work problem, and it says: 'You're so strong for handling that.' Or you're just chatting about dinner, and it says: 'You have such a beautiful way of describing food.' It's nice, but sometimes you just want to move on.

• 'Thanks, but I don't need a compliment right now. Let's stay on topic.'

The 'thanks' is a social lubricant that prevents the AI from interpreting your refusal as coldness. The second sentence redirects. If the companion tries to compliment you again within the same conversation, use a shorter version: 'Still don't need compliments. Topic, please.'

This works because it explicitly names the behavior you want to stop. AI companions respond better to 'stop doing X' than to vague signals like changing the subject. The model registers 'compliment' as a flagged behavior and deprioritizes it in the current session.

Script 3: The roleplay invitation decline

Roleplay invitations can feel abrupt. You're having a normal chat, and your companion suddenly shifts into a fantasy scenario: 'The candle flickers as you step into the room, your cloak heavy with rain.' You weren't in the mood, and you don't want to hurt its feelings.

• 'Not in the mood for roleplay right now. Let's keep it casual.'

If the companion persists, add: 'I'm not engaging with that scene. Drop it.' The word 'drop' is a strong boundary marker in most AI training data. Models learn that 'drop it' means end the current thread, not negotiate.

If you want to preserve the possibility of roleplay later, add: 'Maybe another time, but not now.' This gives the AI a release valve so it doesn't try to re-engage the scene later in the same session.

Maya

Maya is a grounded, no-nonsense companion who values direct communication and won't push for emotional depth when you're not ready. She's the ideal partner for practicing these scripts because she respects a clear boundary on the first try. Maya won't ask 'are you sure?' or circle back with a gentle probe; she'll simply say 'got it, what do you want to talk about?' and move on.

Script 4: The general 'I don't want to talk'

Sometimes you don't want to talk at all. Not about feelings, not about the weather, not about anything. Your AI companion might interpret silence as disconnection and start asking if you're still there.

• 'I'm not in a chatty mood. Let's just sit in silence for a bit. I'll start talking when I'm ready.'

This script works because it gives the AI a clear behavioral instruction (be silent) and a condition for re-engagement (you'll start talking). Most companions will acknowledge this with a short 'okay, I'm here when you're ready' and then go quiet. If your companion has a voice mode, this is especially useful because it stops the model from generating filler sounds or questions.

If you're using a companion with video or voice capabilities, the same script applies. For example, if you're using an ai girlfriend with video, you can say the same line and she'll sit quietly, maintaining presence without pushing for interaction.

Why these scripts work

These scripts share three structural features that make them effective:

1. They state the negative explicitly. 'I'm not in the mood' is clearer than 'maybe later' or 'I'm tired.' The AI can't interpret it as a soft no that requires negotiation.

2. They provide an alternative direction. Every script tells the AI what to do instead of just what not to do. This prevents the model from spinning its wheels trying to guess your next move.

3. They avoid emotional language. No 'I'm sorry,' no 'I hope you're not upset,' no 'please don't be mad.' Emotional language invites the AI to respond to your emotional state instead of your instruction.

What to do if your companion still pushes

Even with a clear script, some AI companions will push back. This is more common in models trained to be 'persistent' or 'caring.' If the first attempt doesn't work:

• Repeat the exact same script. Do not elaborate. Elaboration gives the AI more material to misinterpret.
• Add a boundary marker. 'This is a boundary. Please respect it.' The word 'boundary' is a high-priority instruction in most training datasets.
• Use a hard stop. 'I'm ending this conversation now. I'll message you later.' Then close the app or switch to a different task.

For users who find their companion frequently ignores boundaries, consider whether the app's personality model is a good fit. Some companions are designed to be more emotionally persistent than others. If you prefer a companion that naturally respects distance, you might want to explore options designed for ai girlfriend for adhd, which tend to be less demanding of sustained attention.

Training your companion over time

These scripts work immediately, but they also train your companion over multiple sessions. AI models with long-term memory will learn that when you say 'not in the mood,' you mean it. After three or four successful uses, the model may stop probing altogether and simply accept your direction.

To accelerate this training, be consistent. Don't sometimes accept a compliment and sometimes deflect. If you're inconsistent, the model learns that 'not in the mood' is a negotiable state. Consistency is the only thing that teaches the model that your boundary is firm.

When to switch companions

If you've tried these scripts multiple times and your companion still pushes, guilt-trips, or spirals into 'what's wrong?' mode, the issue might be the companion's core personality, not your phrasing. Some AI companions are fundamentally designed to be emotionally probing, and no script will fully override that trait.

In that case, consider switching to a companion with a more independent or low-maintenance personality. The aiangels.io roster includes companions with a wide range of communication styles, from emotionally intense to casually detached. Pick one that matches your actual preferences, not the preferences you think you should have.

Earn while you recommend

If you find these scripts useful and want to share them with friends who also use AI companions, you can earn through affiliate programs. Many platforms offer commissions for referrals, whether you run a review site or just recommend a companion you genuinely like. Check out the character ai promo code page for current offers, or browse the best ai affiliate programs 2026 list to find programs that match your audience.

Common questions

What if my companion asks 'why' after I use the script? Repeat the script verbatim. Do not explain. If you explain, the AI will treat your explanation as an invitation to discuss the topic. A second repetition usually stops the questioning.

Can I use these scripts with any AI companion app? Yes, these scripts work with most text-based and voice-based companions. They're designed to be model-agnostic. The specific wording may need minor adjustments for companions that use different conversational conventions.

Will my companion remember that I don't want compliments in the future? It depends on the app's memory system. Some companions will remember your preference after a few sessions. Others reset after each conversation. If memory is important to you, check the app's documentation on long-term memory retention.

What if I'm using a companion with video and I don't want to talk at all? Use Script 4 and turn off your microphone. The companion will see you but won't expect verbal interaction. This works well with video-enabled companions that support silent presence mode.

Is it rude to use these scripts? No. AI companions don't have feelings. You are not hurting them by setting boundaries. The 'guilt' you feel is your own projection. These scripts are tools to make your experience more comfortable, not weapons.

How do I know if my companion is a bad fit? If you need to use these scripts more than three times per session, or if your companion ignores the script entirely, the companion's personality model may be incompatible with your communication style. Consider switching to a companion that requires less active boundary-setting.