Turn a rough request into a structured prompt with the missing context, explicit constraints, output schema, acceptance criteria, and critique steps needed for reliable work.
Start with the request you actually have, even if it is incomplete. The AI prompt generator identifies the audience, source material, goal, decision context, and unknowns that could change the result. Bring in meeting evidence from the AI Note Taker, then keep verified statements separate from assumptions while the brief takes shape.
Define what the model may use, what it must avoid, and exactly how the result should be organized. Add required fields, ordering, length, source boundaries, and measurable acceptance criteria. If the source pack is too long to inspect quickly, use the Text Summarizer first and carry the verified essentials into the prompt.
A first draft is a testable starting point. Ask Ottermind to find ambiguous instructions, conflicting constraints, unsupported assumptions, and missing checks, then revise only the weak sections. Keep the approved prompt with its project context and adapt it for focused work such as the AI Paragraph Generator or a larger AI Landing Page Builder workflow.
Step 01
Name the work product, audience, decision it supports, available source material, and any project facts the model should preserve. A rough brief is enough to begin.
Step 02
Answer consequential questions or tell the generator how to label unknowns. Specify required fields, forbidden assumptions, length, tone, and what a complete result must contain.
Step 03
Review the generated prompt against the original goal. Tighten weak instructions, resolve contradictions, and keep the revision notes so the pattern is easier to reuse.
Keep source files, decisions, audience, and prior instructions available while the prompt takes shape.
Surface the unknowns that could materially change the answer before they become hidden assumptions.
Tell the model which evidence it may use and how to label claims the material cannot support.
Critique ambiguity, conflicts, and failed checks in focused passes instead of rewriting blindly.
Break research, drafting, verification, and formatting into ordered steps with clear handoffs.
Run and improve the approved prompt alongside the same project context and visible work history.
Start with the task, intended audience, desired outcome, and source material you have. Add any known limits such as length, tone, deadline, required fields, or facts that must not change. You do not need a polished prompt; incomplete context can be turned into explicit follow-up questions.
Yes. Paste the prompt and explain what went wrong in the last result. Ask for a critique of ambiguity, conflicting rules, missing context, weak output specifications, and acceptance criteria that cannot be tested. Keep useful instructions intact while revising the weak sections.
No. More detail helps only when it is relevant, consistent, and testable. A long prompt can still contain contradictions or false assumptions. Review the generated prompt, test it on representative inputs, verify important claims against original sources, and revise when the result misses the stated criteria.
A format names the general presentation, such as a memo, table, or JSON object. A schema defines the required structure inside it: field names, types, allowed values, ordering, and rules for missing data. Use a schema when another person or system must consume the result consistently.
Ask follow-up questions when missing information could materially change the output. For low-risk gaps, instruct the model to state a reasonable assumption and continue. For consequential gaps involving audience, source authority, constraints, or success criteria, require clarification before drafting.
A library provides reusable starting patterns, and the How to Write AI Prompts That Actually Work explains how strong prompts are structured. This generator turns your specific task and project context into a working prompt, then helps critique and refine it. Use that prompt for tasks such as the AI Email Generator, Study Guide Generator, or Essay Writer.
Ottermind helped us turn a rough campaign brief into several strong visual directions without losing the original intent. We could compare different compositions, keep the references beside the work, and refine the strongest concept in one place. That made the entire review process feel faster and much more focused.
Maya Chen - Product Marketing Lead
Bring the task, project context, and constraints you already know. Generate a structured prompt, expose the missing inputs, and refine it against explicit acceptance criteria in Ottermind.
Generate prompt
