Most people ask AI generic relationship questions and receive generic relationship answers; prompt engineering for couples work means learning to include context (what you've both tried, why it failed, what matters most), ask for specific scenarios rather than principles, and push back when answers feel familiar but disconnected from your actual life.
Prompt engineering is the art of asking AI questions in a way that gets you actual useful answers instead of generic platitudes. In relationships, this is essential—because if you ask ChatGPT "How do I improve my marriage?" you'll get the same advice everyone gets. If you ask it the *right* way, you get personalized insight.
AI models respond to specificity. When you give them context, constraints, and your actual situation, they can give you better thinking. It's like the difference between asking a friend "How do I fix my relationship?" (they'll give generic advice) versus saying "My partner feels like I don't listen during stressful work weeks, but I actually do remember what she says—I just process things slowly. How do I show that I'm engaged without interrupting her?" (now they can actually help).
A weak prompt: "We fight about money. What should we do?"
A strong prompt: "My partner and I have been together 8 years. We earn similar salaries but have different spending styles. She saves 40% of income; I save 15%. We don't have kids. We fight about my discretionary spending, but I feel controlled. I'm not in debt. What's a framework for discussing this that doesn't feel like budgeting blame?"
See the difference? The second gives AI actual material to work with. It understands your values (neither of you is irresponsible), your constraints (no kids, no debt), and your real problem (feeling controlled, not money itself).
When you give AI these three pieces, it can generate thinking that actually feels personalized. The AI still doesn't know your relationship. But it's working with your actual data, not template advice.
AI will still sometimes miss nuance. Use its output as a starting point for your own thinking, not as gospel. The value is in how the prompting forces *you* to clarify what you actually want to know.
Try this: Write out a relationship question you've been sitting with. Now rewrite it using the context-constraint-question formula. Feed both versions to Claude or ChatGPT and compare the responses. You'll immediately see why specificity matters.
Peri can explain this concept, give practical examples, help you decide whether it applies to your situation, or recommend a journey if appropriate.
Explore related journeys or tell Peri what you're working through.