10 ChatGPT Prompts for Therapy That Support Emotional Insight

Therapy is one of the most powerful tools for personal growth and mental health — but the work does not only happen in the therapy room. The right ChatGPT prompts for therapy can support the between-session work: helping you process experiences, articulate emotions, identify patterns, and arrive at your next session with more clarity about what you want to explore.

Important: these prompts are designed to support self-reflection as a complement to professional therapy — not as a replacement. If you are experiencing a mental health crisis, please contact a qualified professional or crisis service immediately.

I have a therapy session coming up. Here is what has been on my mind since my last session: [describe]. Help me: identify the 2-3 themes most important to bring, formulate one clear question I want to explore, and identify any feelings or topics I have been avoiding that might be worth raising. Do not analyze or interpret — help me organize my own thinking.

Why it works: clients who arrive with a clear sense of what they want to explore get more from each hour. The avoidance question is the most valuable — the things we are most reluctant to raise are usually the most worth exploring.

I am feeling something I am having difficulty naming. Let me describe it: [describe — physical sensations, thoughts that accompany it, what triggered it, how long it has been there]. Offer 5 to 8 emotion words or phrases that might fit, explain what each means and how it differs from the others, and ask one question that might help me narrow it down further.

Why it works: naming emotions with precision reduces their intensity and improves regulation. The difference between feeling ‘sad’ and ‘grief-adjacent’ is not just semantic: precision creates distance and understanding where vagueness creates overwhelm.

Help me work through a CBT-style thought record. The situation: [describe]. My automatic thought: [describe]. My emotional response: [emotion and intensity 0-10]. Help me examine this thought by asking: what evidence supports it, what contradicts it, what cognitive distortions might be present, what a more balanced thought might be, and how I feel after considering this alternative (0-10).

Why it works: CBT thought records interrupt the automatic thought-emotion loop with structured examination. The before-and-after intensity rating makes the impact of the exercise concrete — which reinforces the practice and builds evidence that the technique actually helps.

I want to understand my attachment patterns. I will describe how I tend to behave in close relationships: [describe — how you respond to intimacy, conflict, distance, and abandonment]. Based on this, help me understand: which attachment style this resembles, where it likely developed, how it shows up in relationships in ways I might not recognize, and one thing I could practice to develop more secure relating.

Why it works: understanding the developmental origin of an attachment pattern makes it feel less like a character flaw and more like an understandable response to early experience — which is what creates the psychological space to change it.

I am processing a loss: [describe]. Generate 8 journaling prompts that help me move through grief without rushing it. The prompts should: create space for the full range of grief emotions including anger, relief, and guilt alongside sadness, help me honor what was lost, allow me to examine what I fear the loss means for my future, and gently invite me to consider what I am carrying forward. Avoid prompts that rush toward acceptance.

Why it works: the full emotional range instruction and ‘avoid rushing acceptance’ directive produce prompts that honor the actual experience of grief rather than the socially acceptable version of it.

I need to set a boundary with [describe the person and relationship]. The specific behavior I want to address: [describe]. What I need instead: [describe]. Write 3 versions: a gentle version for someone I want to preserve the relationship with, a direct version for someone who has ignored softer approaches, and a script for what to say if they push back or challenge the boundary.

Why it works: most boundary-setting fails at the pushback stage. Having the response to ‘but why?’ prepared in advance is what makes the boundary stick rather than collapsing under relational pressure in the moment.

I want to examine a core belief: [state the belief, e.g., ‘I am not enough’]. Help me explore it by: tracing where it likely came from, examining the evidence I use to confirm it while ignoring contradicting evidence, understanding what function it serves — what it protects me from — and developing an alternative belief that is more accurate and still feels honest rather than just positive. Do not dismiss the belief — help me understand it.

Why it works: core beliefs resist change because they feel true. Understanding the selective attention that maintains them — and the protective function they serve — creates the psychological distance needed to begin questioning them.

I just experienced an anxiety episode about [describe]. Help me debrief it: [describe what happened]. Examine: what the anxiety was trying to protect me from, whether the threat was real or anticipated, what my body was telling me, what I did that helped or made it worse, and one thing I can do before the next similar situation. Do not minimize the anxiety — help me understand what it was communicating.

Why it works: treating anxiety as a communication rather than a malfunction reduces shame and increases insight. The ‘what was my body telling me’ question connects somatic experience to emotional meaning — where the most useful information often lives.

I am struggling to be compassionate toward myself about [describe a situation where you are being self-critical]. Help me practice self-compassion in three steps: acknowledge what I am feeling without minimizing it; remind me that struggle and imperfection are universal human experiences with a specific example; and help me write what I would say to a close friend in exactly the same situation, then apply that same kindness to myself.

Why it works: this follows Kristin Neff’s three-component model of self-compassion: mindfulness, common humanity, and self-kindness. The friend framing is the most powerful element — people are consistently kinder to friends in the same situation than to themselves.

I just finished a therapy session. Here is what we talked about: [describe the main themes and insights]. Help me: identify the most important insight in one sentence I can return to, connect it to a pattern we have explored before if I describe one, identify one small concrete action I can take this week that reflects today’s insight, and note one thing I want to bring back to my next session.

Why it works: therapeutic insights that are not acted on between sessions often dissolve before the next appointment. The one-sentence distillation, concrete action, and next-session note create the structure that turns a session into ongoing growth.

The most effective ChatGPT prompts for therapy-adjacent work require honesty and specificity. If you describe situations vaguely or present the acceptable version rather than the true one, the reflection will be shallow. Approach these prompts with the same honesty you bring to your best therapy sessions. Always share significant insights with your therapist — these prompts work best as a bridge between sessions, not a replacement.

Different AI models bring different conversational qualities to emotionally sensitive topics. Chat Smith gives you access to Claude, GPT, Gemini, Grok, and DeepSeek in one platform — so you can use Claude for the most nuanced, emotionally attuned reflective conversations, GPT for structured frameworks like thought records and post-session integration, and Gemini when you want a different perspective on the same emotional material. For sensitive personal work, Claude’s careful, non-judgmental tone tends to produce the most therapeutically useful responses.

Chat Smith also lets you save your most-used prompts as reusable templates. Store your pre-session preparation prompt and post-session integration tool so they are available consistently — building a between-session practice that deepens the therapeutic work you are doing with your professional.

Therapy works best when the insights generated in the room are extended into the rest of your life. The prompts in this guide give you a structured, thoughtful way to do that between-session work. For the multi-model AI platform that supports that practice with care and nuance, Chat Smith is built for exactly that. If you are in crisis, please reach out to a qualified mental health professional or crisis service.

1. Can ChatGPT replace therapy?

No — and it is important to be clear about this. Therapy involves a trained professional who can hold therapeutic context across sessions, notice what you are not saying, work with trauma safely, and provide clinical assessment and treatment. ChatGPT is a self-reflection tool that can support the between-session work of therapy. It cannot diagnose, treat, or provide the relational safety that makes therapy transformative. If you are struggling with your mental health, please seek professional support.

2. Is it safe to share personal emotional content with AI?

You control what you share. You do not need to use real names or identify specific people to make these prompts useful — describing the dynamic or the emotion is sufficient. For very sensitive content involving trauma, abuse, or crisis, please work with a qualified professional rather than an AI tool.

3. Should I share the outputs from these prompts with my therapist?

Yes — if they feel relevant. Sharing what came up in a between-session reflection — an insight, a feeling you named, a pattern you noticed — can enrich your therapy sessions significantly. The between-session thinking you do with these tools is material your therapist can work with.