Writing effective AI prompts follows a structure: role, task, format, and constraints. The role sets the AI's perspective (act as a senior copywriter). The task defines what to produce (write a 3-email welcome sequence). The format specifies the output shape (bullet list, table, JSON). Constraints narrow the scope (no jargon, B2B audience, 150 words each). This 4-part structure works across ChatGPT, Claude, Gemini, and any other large language model. At aidowith.me, the Practical Prompts route covers 15 steps in about 1 hour 15 minutes. You build 3 reusable prompt templates during the route: one for writing tasks, one for analysis, and one for structured data output. Each template includes fill-in fields so you can adapt it in 60 seconds for any new task. You finish with a personal prompt library, not just notes. Over 3,000 professionals have used this route as their entry point for AI tools at work.
Last updated: April 2026
The Problem and the Fix
Without a route
- Vague prompts produce vague outputs. Without a structure, you get generic results and spend 30 minutes rewriting.
- Most guides list principles but do not give you a working template to fill in right now.
- Each AI tool feels different, but one prompt structure works across all of them when applied correctly.
With aidowith.me
- Apply the 4-part prompt structure: role, task, format, and constraints for consistent outputs every time.
- Build 3 reusable prompt templates during the route and adapt them for any task in under 60 seconds.
- Get the same quality results across ChatGPT, Claude, and Gemini using one transferable framework.
Who Builds This With AI
Marketers
Content, campaigns, and briefs done in hours instead of days.
Sales & BizDev
Prep calls, draft outreach, research prospects in minutes.
Managers & Leads
Reports, presentations, and team comms handled faster.
How It Works
Write the role and task
Open ChatGPT and write: 'Act as [role]. Your task is to [specific output].' Test it with one real work task you need to finish today.
Add format and constraints
Extend your prompt: 'Format the output as [structure]. Keep each section under [word count]. Avoid [specific elements].' Run it and compare the output to your first attempt.
Save it as a reusable template
Replace the specific task details with placeholders: [role], [task], [audience], [format]. Save the template in a note or doc so you can fill it in for the next task in seconds.
Build Your Personal Prompt Library
The aidowith.me Practical Prompts route walks you through 15 steps and ends with 3 reusable templates. Start at so.aidowith.me.
Start This Route →What You Walk Away With
Write the role and task
Add format and constraints
Save it as a reusable template
Get the same quality results across ChatGPT, Claude, and Gemini using one transferable framework.
"The 4-part structure changed everything. I went from getting mediocre AI outputs to getting drafts I can use without rewrites."- Content Strategist, marketing agency
Questions
The most reliable approach is the role-task-format-constraints structure. Start every prompt with the AI's role (act as a senior marketer), followed by a specific task (write a LinkedIn post), then the format (3 short paragraphs, no hashtags), and any constraints (for a B2B SaaS audience, tone is professional but direct). This structure works across all major AI tools and gets usable outputs on the first attempt far more often than open-ended prompts do.
The same prompt structure works across ChatGPT, Claude, and Gemini, though each model has different default tendencies. ChatGPT is more conversational. Claude tends toward longer, more structured outputs. Gemini integrates well with Google Workspace tasks. The role-task-format-constraints framework transfers across all three. The aidowith.me Practical Prompts route includes examples tested on multiple models so you can adapt without rewriting from scratch.
Prompt length should match task complexity. A short copywriting task needs 30 to 50 words. A complex analysis or multi-section document needs 100 to 200 words with clear section headers. The biggest mistake is writing too little context, not too much. If the AI asks follow-up questions or produces a generic result, the prompt is missing constraints or context. Add one specific detail at a time until the output matches your expectation.