The quality of what AI gives you depends entirely on how specifically you ask; vague prompts get generic responses, while precise ones unlock genuinely useful thinking. Learning to engineer prompts means treating the AI as a thinking partner you're directing, not a magic box you're hoping will solve your problem.
You've probably asked an AI something and gotten a generic, unhelpful answer. Then you asked slightly differently and got something actually useful. The difference isn't the AI—it's the prompt. A prompt is just the question or instruction you give an AI. Prompt engineering is the skill of asking in a way that gets the best response.
In relationships and communication, prompt engineering is crucial because relationship advice that ignores your specific situation is worthless. "Talk to them about it" isn't helpful if you don't know how. But "My partner gets defensive when I use the word 'always.' How can I address a pattern of behavior without triggering that defensiveness?" opens the door to actual useful guidance.
Think of an AI like a search engine. Bad search: "relationship." Good search: "How do I tell my partner I need more time alone without them feeling rejected?" The same principle applies to AI. The more specific, detailed, and clear your prompt, the more targeted and useful the response.
A weak prompt: "How do I fix my communication?" A strong prompt: "I tend to get quiet when I'm hurt instead of saying what's wrong. My partner interprets silence as anger or distance. How can I start speaking up without becoming confrontational?" The second prompt tells the AI what the actual problem is, what pattern you're stuck in, and what you're trying to achieve. It can give you advice that's actually relevant to you.
First, provide context: relationship type (partner, friend, family), relationship length, relevant history. Second, name the specific problem: not "we fight," but "we disagree about finances and conversations turn accusatory." Third, include constraints: "I want to set a boundary but I don't want them to feel attacked." Fourth, tell the AI what kind of response helps: "Give me specific wording for how to start this conversation."
The best prompts are almost like bringing the AI into your confidence. You're saying: "Here's my situation, here's what I've tried, here's what I'm worried will happen, here's what I'm hoping for." With that information, the AI can actually help.
Asking AI to read minds: "Is my partner cheating?" (AI can't know.) Asking without context: "Should I break up?" (AI needs to know ages, relationship length, what's happening.) Being too vague about emotions: "They made me mad." (The AI needs to know what specifically triggered you.) Hiding important details: "Should I tell them?" (Tell them what? The context changes everything.)
Prompt engineering isn't a fixed skill. It's a conversation. You ask, see what the AI gives you, refine. If the first response is generic, follow up: "That's a good start, but my partner specifically gets defensive when..." The AI learns from your clarifications and gets better.
Try this: Think of a relationship challenge you're facing. Write two versions: a vague one ("How do I communicate better?") and a detailed one that includes context, the specific problem, and what you're hoping for. Ask both versions to the same AI and compare responses. Notice how much more useful the detailed prompt's answer is.
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.