Becoming an AI Power User

6.3Workflow Redesign, Creativity, and Growth

They create personalized software, thoughtful gifts, and solutions tailored to exactly how they live and work

Developing

A specific need can guide a small, useful build

A personal reading notebook might need only a book title, a question, and a way to find that question later. AI can help you specify and build that small tool, or adapt an existing one. The opportunity comes from knowing whose need you are serving: you may build a feature useful to one person even when it would not justify a general product.

Custom software still needs testing and upkeep. A can help diagnose faults and implement changes, but you need a way to check those changes and recover saved data. A simple reading notebook has fewer dependencies than a dashboard connected to several changing services. Include that difference when deciding what to build.

Choose something you wish existed and describe what would make it useful or meaningful to its intended recipient. Ask AI to scope a first version that demonstrates that value within the time you have, with a smaller fallback if needed.

A reader uses a small personal reading tool for the book, a question and finding it later, alongside a simple paper equivalent and maintenance tool. — When building gets cheap, you may build directly for very specific needs.

The person and purpose tell you which details matter. A reading notebook that keeps your unanswered questions beside each book may suit you better than a feature-rich library app. A gift might be a recipe collection organized around meals you shared with a friend. AI can help with design and production while you provide the context that distinguishes this project from a generic one. Compare a custom build with adapting an existing tool before taking on its upkeep.

They build personalized versions of the tools they use

Suppose you keep notes like this: 'The Left Hand of Darkness: How does the narrator revise his assumptions? Return to chapter seven.' You want to find unanswered questions across books. A useful first request is, 'Build a reading notebook with fields for book, question, reference, and answered or open. Let me filter for open questions and save my notes.' This hypothetical example gives the agent concrete behavior to build and a small result for inspection.

An agent can also help discover the requirement. Show it examples of the notes you keep, or ask it to interview you about what is awkward in your current reading app. If it proposes a rating system you would never use, explain why; if it proposes a useful connection between notes, try it. The agent contributes design ideas while your reactions provide further evidence about what fits.

Test the reading notebook by adding the sample question, marking it answered, and filtering for open questions. Then close and reopen the notebook to check that the note is still there. Ask the agent to help verify those behaviors and provide a way to export the notes. A screen that looks right is only part of the result; the saved information must behave as promised.

They use AI to make thoughtful gifts and creative projects for people they care about

One of the overlooked uses of AI is creative personalization. A thoughtful gift is one where the giver clearly spent time understanding what the recipient values. AI can help you express that care in a form you would otherwise struggle to make.

For example, a parent could use AI to create a personalized children's story featuring their child's interests, friends, and favorite places. A friend uses AI to design a custom recipe book with the specific dishes they have cooked together over years, complete with the stories behind each one. Someone creates a personalized travel guide for a friend's upcoming trip, drawing on their specific interests, pace preferences, and dietary needs.

For the recipe book, bring the dishes you cooked together and any stories you want to include. AI can suggest an order, draft introductions, or propose illustrations; judge those ideas against what you know of your friend. Check names, occasions, and recipe details against the material you supplied. A proposed detail may inspire you, but ask the assistant to label an imagined scene so it cannot pass as a shared memory.

Ask permission before using AI to imitate someone. A recipe collection can express care through dishes and memories you are entitled to share. A video that imitates your friend's voice or likeness calls for their explicit permission for that use. Their possible enjoyment does not establish consent. When a project uses private memories or personal details, check whether the intended audience and use fit the permission you have.

Personal tools can address needs existing tools serve poorly

Some combinations of needs are poorly served by the tools available to you. A freelancer who needs to track which client conversations mentioned specific regulatory requirements. A researcher who wants to cross-reference research notes and flag contradictions between sources. A team lead who needs a weekly summary that brings project status, customer issues, and staffing constraints into a report they can scan before Monday.

AI can help with projects whose value is local: a report for one team, a notebook for one reader, or a reminder that follows a particular routine. Ask it to review examples of the task and suggest a small solution. Check whether an existing feature, template, or simple change in your routine already meets the need before building new software.

A gives you candidates, and a workflow map shows what a proposed tool must read and produce. Add the creative question: what would be enjoyable or useful to make? You may do a playful project for its own sake, even when saving time is not the point.

Try a manageable version and let use guide the next change

A large scope can make a creative project difficult to finish. A small first version gives you something to try while the design is still easy to change. A personalized dashboard does not need to be perfect on day one. It needs to show the indicators that help you make the decision it supports, in an easily understood format. A custom gift does not need to be a polished product. It needs to be thoughtful and complete enough to give.

Use the time you have as a planning constraint, and leave room to test the result. Describe the goal, ask for the smallest version that demonstrates it, and try that version. In the reading notebook, using the open-question filter may show whether a new field would help. Let what you learn guide the next change. Play can suggest the next experiment too: perhaps it would be fun to collect questions shared across several books.

When an approach works and you expect to use it again, ask AI to extract the useful instructions and checks. The reading-notebook project might leave a reusable requirement to preserve user-entered notes after reopening and provide an export. Test that requirement in the next relevant build, adapting it to the project. Later requests can reference the established standard without rediscovering the same requirement each time.

Build something suited to a personal need or interest

Plan a personalized project around what its recipient needs

Claude helps you define the simplest first version of something you wish existed. You decide what to build and review a plan sized to the time you have.

Copy the promptUse the copy button above the box.
Paste it in Claude or CodexMake sure your active project is enabled.
Answer the questionsGive clear answers so it understands what you want made.
Test it to see if it worksTry the output and refine until it does what you need.

Problem: A general-purpose tool or gift may miss the details that matter to a particular person. Intent: I want to plan a personalized tool, creative gift, or solution whose value I can test. Ask what I wish existed, who it is for, and what would make it useful or meaningful. Ask how much time I have, which tools I can use, and how much upkeep I am willing to take on. Use short follow-up exchanges to clarify details that affect the plan. Ask for relevant examples, notes, or preferences I can provide or point you to. If you cannot access that material, ask what I can paste or connect. Distinguish supplied facts and preferences from your design suggestions. Do not invent personal memories, writing samples, or preferences for me or the recipient. 1. Explain the useful or personal value the first version should demonstrate and the capabilities it needs. For a tool, consider whether adapting something I already use would meet the need. 2. Separate essential behavior from features and polish that can wait. Leave time to test the first version and suggest a smaller fallback if the plan exceeds my available time. 3. Recommend an approach using the tools available to me: chat for planning and text, a coding agent for software, image generation for visuals, or a suitable combination. Identify missing access or unverified tool requirements. 4. Write the steps to build and test the first version. Label uncertain timings. For software that stores information, include checks that saved data survives reopening and can be exported or recovered. 5. Include checks for fit: does the result serve the intended person and use the supplied details accurately? If a project would imitate another person's voice or likeness, or use their private material, identify the permission needed for the intended use. 6. Suggest what to observe while using the first version. Base later changes on those observations and ideas I want to explore. Scope: plan and scope only. Do not start building until I approve the plan. Do not save a personal profile, voice, or preferences as part of this exercise. If I later request such an asset, show the findings supported by my actual material and wait for my confirmation before saving it.

When the plan fits, give it to the assistant that will build the project and identify the or material it may use. Ask it to create the first version and carry out the agreed checks. For the reading notebook, confirm that notes survive reopening and can be exported; for a gift, check the details against the memories or examples you supplied.