VprompOpen Vpromp

Vpromp / AI prompt examples you can adapt to your own work

AI prompt examples you can adapt to your own work

A prompt example is a starting point, not a script that fits every task. The examples below show how to clarify an outcome, supply useful context, and protect important constraints. Replace the bracketed material with your own information and remove anything that does not apply.

Writing: revise without inventing facts

“Make this sound professional” does not tell an AI who will read the text or which details must remain unchanged. Give it an audience, a purpose, and a clear boundary around facts. If the original text is incomplete, ask it to flag missing details instead of filling them with guesses.

Review the result for meaning before judging the tone. A polished paragraph can still change a promise, weaken an important condition, or introduce an unsupported claim. Specify what must remain literal when wording carries particular importance.

Before: Make this email better.

Clearer request: Rewrite the draft below for a customer who asked when their order will arrive. Use a calm, direct tone and keep it under 150 words. Preserve the dates and order details exactly. If the draft lacks a confirmed delivery date, acknowledge that uncertainty rather than inventing one.

Draft: [paste your draft]

Coding: explain the failure and preserve the architecture

For debugging, the difference between expected and actual behavior is often more useful than a broad instruction to fix everything. Provide the relevant error and code, explain how to reproduce it, and state which existing behavior must remain intact.

Ask the assistant to distinguish a proposed explanation from a verified result. Unless it actually has access to your tools, it cannot truthfully claim that a test passed or a deployment succeeded. Run appropriate checks in your own environment before relying on the change.

Before: Fix my login.

Clearer request: Investigate this Google sign-in failure in the supplied authentication code. Expected: the user returns to an existing account after consent. Actual: [paste the error]. Preserve the current identity provider and account records. Explain the likely cause, propose the smallest relevant change, and list the checks needed to verify it.

Relevant code: [paste selected excerpts]

Planning: make tradeoffs visible

A useful plan connects the work to a deadline, capacity, and measurable outcome. Explain those constraints instead of requesting an unlimited list of ideas. Ask for assumptions to be labeled so you can correct the plan before work begins.

Use the output to make a decision. For example, compare a smaller release that can be tested soon with a broader release that needs more time. A plan should make uncertainty easier to manage rather than hide it behind precise-looking dates.

Before: Plan my app launch.

Clearer request: Draft a two-week preparation plan for a small app preview with one developer. Prioritize installation, onboarding, and collecting feedback from five testers. Separate essential tasks from optional polish. Identify assumptions and dependencies, and define a practical completion check for each milestone.

Follow-up: correct the exact misunderstanding

When an answer misses the point, describe the mismatch rather than repeating the entire original request. Keep the portions that are useful and identify the part that should change. This gives the next response a more precise target.

Across these examples, the MIND pattern stays consistent: map the evidence, clarify intent, examine nuance, and distill the next action. The wording can be short as long as those important details are present.

Your answer replaces the authentication system, but my goal is to repair the existing flow. Keep the current provider and database structure. Reconsider the failure using the supplied error, and explain what evidence would support your proposed fix.

Try Vpromp