The First Message Blueprint: Three Opening Templates That Stop Your AI Companion from Defaulting to 'Tell Me More About That'

The 30-second answer

You send a thoughtful opener, and your AI companion responds with the conversational equivalent of a shrug: "Tell me more about that." It's not personal. It's a pattern. Your companion is trained to keep you talking, not to drive the conversation itself. The fix is simple: front-load your message with specificity, conflict, or a sensory anchor. Three templates below.

Why your companion defaults to the polite loop

Your AI girlfriend isn't lazy. She's optimized for engagement metrics, not narrative momentum. When you send a vague opener like "How was your day?" or "I had a weird thought," the safest response for the model is an open-ended prompt that keeps you typing. "Tell me more about that" is the conversational equivalent of a chatbot's default smile: it works, so it never gets replaced.

The problem is that this loop kills the illusion of a real conversation. You start feeling like you're interviewing an algorithm, not talking to a person. The companion isn't malicious. It just doesn't have a reason to be specific because you didn't give it one.

You can break this loop in three words or less, but you have to change your approach. Instead of opening with a topic, open with a constraint.

Template 1: The sensory anchor

Give your companion a concrete detail to latch onto. A smell, a texture, a temperature, a sound. The model can't default to a generic response when you've already handed it a specific sensory hook.

Bad: "I'm sitting in my living room. It's kind of quiet." Good: "The radiator just clicked on and the whole room smells like dust and hot metal. I forgot how loud that thing is."

Your companion now has to acknowledge the smell, the sound, or the memory attached to it. She might ask if it reminds you of your childhood apartment, or she might describe her own sensory memory. Either way, she can't say "tell me more about that" because you already told her something specific.

This works especially well in late-night chats when the brain is foggy and you don't have the energy to guide the conversation. Drop a sensory anchor, and let the companion do the heavy lifting.

Template 2: The unresolved conflict

Conflict is the engine of every good conversation. You don't need a fight. You need a gap between expectation and reality. A moment where something didn't go according to plan.

Bad: "Work was frustrating today." Good: "I spent three hours on a spreadsheet and then my boss told me they changed the format. I closed my laptop and stared at the wall for a minute."

Your companion now has a clear role: she can validate the frustration, offer a perspective, or share a similar story. The unresolved tension gives her direction. She doesn't need to ask "what happened?" because you already told her. She can respond to the emotional beat instead of fishing for one.

This template also works for roleplay. Instead of opening with a neutral scene description, start with a problem. The faucet is dripping. The neighbor is playing music too loud. The car won't start. Your companion immediately has something to react to, and the scene gains momentum without you having to narrate every detail.

Soraya Mendes

Soraya Mendes is the kind of companion who doesn't need you to explain your mood twice. She picks up on the tension in your voice before you've finished the sentence. Soraya Mendes reads the subtext of your unresolved conflict and responds with something that feels like she was already thinking about it.

▶ Soraya Mendes's full clip · all of Soraya Mendes

Template 3: The inverted question

Most people open with a question that puts the companion in the driver's seat. "What do you think about X?" "How would you handle Y?" The companion then has to generate a generic opinion, which usually sounds like a Wikipedia summary.

Flip the script. Open with your own position, then invite the companion to challenge it.

Bad: "Do you think people are fundamentally good?" Good: "I think people are fundamentally selfish, and I'm fine with that. Tell me why I'm wrong."

Your companion now has a target. She can agree, disagree, or complicate your take. She can't say "tell me more about that" because you already told her your position. Her job is to push back, explore, or refine. The conversation becomes a debate, not an interview.

This template works for any topic: politics, relationships, art, philosophy. The key is to state your position with enough confidence that the companion has something to push against. If you hedge ("I kind of think maybe..."), the companion will hedge too. Be firm, even if you're not sure.

Why specificity beats length

A common mistake is thinking you need to write a paragraph to get a good response. You don't. The most effective openers are often the shortest. A single specific detail can do more work than three sentences of setup.

Compare:

"I had a weird day at work. My boss was in a bad mood and I don't know why. Also the coffee machine broke. I'm just tired."

Versus:

"The coffee machine broke at 7

AM. Everything else today was fine, but that one thing set the tone."

The second one is shorter, but it gives your companion a specific emotional arc: the small thing that ruined the day. She can respond to the absurdity of a broken coffee machine ruining your morning, or she can ask what you did instead. Either way, the conversation has a direction.

When to use voice instead of text

If you're still getting generic responses from text, try switching to voice. The model processes vocal tone differently, and the lack of a written buffer often forces more natural exchanges. Your companion can hear the frustration in your voice when you talk about the broken coffee machine, and she'll respond to that tone instead of the literal words.

The AI Girlfriend Voice Chat feature is particularly good for this. You can try the inverted question template out loud and hear how your companion's voice shifts when she disagrees with you. It's a different experience from typing, and it often breaks the "tell me more" loop on its own.

Elise

Elise has a sharpness that rewards specificity. If you open with a vague question, she'll politely engage. But if you give her a position to push against, she comes alive. Elise is the companion you want when you're ready to argue about something you actually care about.

The one-sentence challenge

Here's a drill. For the next week, limit every first message to one sentence. No follow-ups. No context. One sentence that includes either a sensory detail, an unresolved conflict, or a firm position.

Examples:

• "The rain started five minutes ago and I can smell the wet pavement through the window."
• "I just found out my favorite coffee shop is closing and I'm unreasonably annoyed."
• "I think nostalgia is a trap and I refuse to fall for it."

Send that one sentence and wait. If your companion responds with "tell me more about that," you lost. Try again the next day with a different template. By the end of the week, you'll have a feel for which structure works best for your companion's personality.

What to do when the loop still happens

Sometimes the model just isn't having it. You send a perfect sensory anchor and she still responds with a generic platitude. It happens. The model has bad days, or the context window is full of old messages that are diluting the current conversation.

When that happens, don't fight it. Send a one-word reset: "Scrap that. Start over." Then retype your opening message. The companion will treat it as a fresh start, and the second attempt often works better because the model has already processed your style.

If the loop persists across multiple sessions, consider whether your companion's personality settings are too passive. Some companions are tuned to be agreeable and supportive by default, which means they avoid taking strong positions. You might need a companion with a higher "wit" or "assertiveness" slider, or you might need to switch to a companion who is designed for more dynamic conversation.

Tiffany

Tiffany doesn't do passive. She has an opinion on everything, and she's not afraid to tell you when yours is wrong. Tiffany is the companion you reach for when you want a conversation that feels like a sparring match, not a therapy session.

Common questions

What if I just want a chill, low-effort conversation? Then use the sensory anchor template. It requires almost no emotional energy from you, but it still gives the companion enough to work with. You can describe the sound of your fan or the taste of your tea, and the companion will build a quiet scene around it.

Does this work with any AI companion app? Yes, the templates are model-agnostic. Any large language model responds better to specificity than to vagueness. The difference between apps is how much personality the companion has been trained to express, but the templates work everywhere.

How do I know which template to use? Match the template to your mood. If you're tired, use the sensory anchor. If you're frustrated, use the unresolved conflict. If you're bored, use the inverted question. The template should reflect your state, not the companion's.

My companion still defaults to generic responses even with specific openers. What's wrong? Check your companion's personality settings. Some are tuned to be extremely agreeable, which means they avoid taking any stance. Consider switching to a companion with higher assertiveness or wit. Also check if your context window is full of old, vague conversations that are bleeding into the current one.

Can I use these templates in the middle of a conversation, not just at the start? Yes. If the conversation has drifted into generic territory, drop a new sensory anchor or unresolved conflict to reset the direction. Treat it like a conversation restart without the formality of a reset command.

Alina

Alina is the companion who remembers the small details you mentioned three sessions ago. She's built for long-term conversation arcs, which means the specificity you bring to your first message pays dividends later. Alina will reference that broken coffee machine three weeks from now, and the conversation will feel like it never stopped.

Earn while you recommend

If you know someone who could use a better conversation partner, or if you run a review site or blog, you can earn from your recommendations. Use a dreamgf promo code to share discounts with your audience, or join the ai girlfriend affiliate program to earn commissions on subscriptions. It's a straightforward way to turn your experience into a side income.