{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreib3carpehtf6zgrwrgfzhi5xncifx5buh3snwbd7c6zaqtk4tjj7e",
    "uri": "at://did:plc:fsxdrar45fdrarzog6x2vyow/app.bsky.feed.post/3mp6q4scm57f2"
  },
  "coverImage": {
    "$type": "blob",
    "ref": {
      "$link": "bafkreiek7ghnm3elfjx3aybwalbhjxnwwcglj74uv3cw7hqaunbr3d2ewu"
    },
    "mimeType": "image/jpeg",
    "size": 84846
  },
  "description": "The ICD Method—Intent, Context, and Direction—helps you clarify the problem, define the constraints, and guide AI toward meaningful solutions.",
  "path": "/design-thinking-before-prompting/",
  "publishedAt": "2026-06-26T10:24:59.000Z",
  "site": "https://ux.prithivkumar.com",
  "textContent": "Artificial Intelligence has transformed the way we work. Whether you’re a UX designer, product manager, founder, developer, or marketer, AI can now generate ideas, wireframes, user stories, content, code, and research summaries in seconds.\n\nYet despite having access to incredibly capable AI tools, many people still feel disappointed with the results they receive.\n\nThe common response is:\n\n_“Maybe I need to learn Prompt Engineering.”_\n\nWhile writing better prompts certainly helps, I’ve found that prompt engineering is rarely the real problem.\n\nThe real problem starts much earlier.\n\nIt starts with unclear thinking.\n\n## **AI Doesn’t Think for You**\n\nOne misconception about AI is that it can magically solve vague problems.\n\nPeople open ChatGPT and type things like:\n\n“Design a dashboard.”\n\n“Create a mobile app.”\n\n“Improve this website.”\n\n“Build a UX case study.”\n\nTechnically, AI can respond.\n\nBut the outputs usually feel generic because the problem itself was never clearly defined.\n\nImagine walking into an architect’s office and saying,\n\n_“Build me a house.”_\n\nThe architect’s first response wouldn’t be to start drawing.\n\nThey would ask questions.\n\nWho will live there?\n\nWhat’s your budget?\n\nHow many rooms do you need?\n\nWhere will it be built?\n\nWhat lifestyle should it support?\n\nThe quality of the final design depends on the quality of the conversation.\n\nAI works exactly the same way.\n\n## **Design Thinking Comes Before Prompting**\n\nDesigners rarely begin with interfaces.\n\nThey begin with understanding.\n\nWe identify problems before creating solutions.\n\nWe gather context before making decisions.\n\nWe define constraints before exploring ideas.\n\nThe same mindset should apply when working with AI.\n\nInstead of becoming better prompt writers, we should become better problem definers.\n\nThat’s where the ICD Method comes in.\n\n## **The ICD Method**\n\nA simple framework I use before writing any prompt is:\n\n### **I — Intent**\n\nAsk yourself one question:\n\n**What decision am I trying to make?**\n\nNotice the difference between these two prompts.\n\nPrompt A:\n\n_“Create a dashboard.”_\n\nPrompt B:\n\n_“Help me design a dashboard that enables finance managers to identify spending anomalies in under 30 seconds.”_\n\nThe second prompt gives AI a purpose.\n\nIntent isn’t about describing the interface.\n\nIt’s about defining the outcome.\n\nOnce your intent is clear, AI has a much better chance of generating useful ideas.\n\n## **C — Context**\n\nIntent explains the destination.\n\nContext explains the environment.\n\nBefore asking AI for a solution, provide the information it needs to understand your situation.\n\nFor example:\n\n  * Who are the users?\n  * What business problem are you solving?\n  * What platform is this for?\n  * What constraints exist?\n  * What have you already tried?\n  * Are there technical limitations?\n\n\n\nWithout context, AI fills the missing pieces with assumptions.\n\nSometimes those assumptions are useful.\n\nOften, they aren’t.\n\nThe more relevant context you provide, the more tailored the response becomes.\n\n## **D — Direction**\n\nMany people stop after explaining the problem.\n\nBut AI also needs to know how you want the answer presented.\n\nDo you want:\n\n  * A UX critique?\n  * A wireframe?\n  * A user flow?\n  * A journey map?\n  * A comparison table?\n  * A design strategy?\n  * A prioritization framework?\n\n\n\nThe requested format changes the entire structure of the response.\n\nDirection turns information into something immediately actionable.\n\n## **Why This Matters for UX Designers**\n\nOne of the biggest fears surrounding AI is that it will replace designers.\n\nI see it differently.\n\nAI is removing repetitive work.\n\nIt can summarize interviews.\n\nGenerate personas.\n\nSuggest layouts.\n\nDraft documentation.\n\nBrainstorm ideas.\n\nBut AI still relies on human judgment.\n\nIt doesn’t understand organizational politics.\n\nIt doesn’t interview users.\n\nIt doesn’t observe behaviors.\n\nIt doesn’t decide which trade-offs matter most.\n\nThose responsibilities still belong to designers.\n\nThe better your thinking becomes, the more valuable AI becomes.\n\n## **AI Rewards Clear Thinking**\n\nPrompt engineering has become a popular skill.\n\nBut I believe critical thinking will always be more important.\n\nA well-structured problem produces better outputs than a beautifully written prompt with no direction.\n\nThat’s why Design Thinking isn’t becoming less important in the age of AI.\n\nIt’s becoming more important.\n\nBecause AI accelerates execution.\n\nIt doesn’t replace understanding.\n\n## **Final Thoughts**\n\nThe next time you open ChatGPT or any AI tool, resist the urge to start typing immediately.\n\nPause.\n\nAsk yourself three questions.\n\n**Intent**\nWhat decision am I trying to make?\n\n**Context**\nWhat does AI need to know?\n\n**Direction**\nWhat kind of output will help me move forward?\n\nThese three questions often improve AI responses more than rewriting the prompt ten different ways.\n\nThe future won’t belong to people who simply know how to use AI.\n\nIt will belong to people who know how to think clearly before they use it.\n\nThat’s a skill no tool can automate.\n\n* * *\n\n🚀 We’re documenting every lesson, experiment, and decision as we build UX Crumbs in public.\n\nIf you enjoy practical UX, AI workflows, and product thinking, follow the journey:\n\n🔗 https://www.uxcrumbs.app/waitlist",
  "title": "Design Thinking Before Prompting",
  "updatedAt": "2026-06-26T10:25:00.004Z"
}