
Try a Real Prompt with ChatGP
Try a real prompt with ChatGPT
Build a structured RCTOF prompt, copy it into ChatGPT, and get actionable M&E outputs.
Your mission: Use the RCTOF framework (Role, Context, Task, Output, Filter) to craft a precise AI prompt. As you fill in the fields below, your prompt is generated automatically. Then copy & paste into ChatGPT / Claude.
RCTOF Framework: Practical Prompt Structure
Five-component architecture aligned with the 4Cs (Clear, Complete, Contextual, Constrained).
Expert identity
Sector, country, donor
Specific action
Format & structure
Constraints, ethics
Show complete RCTOF example (girls' education)
C – Context: Rural Kenya, adolescent girls (12-18) at risk of dropout. Donor requires gender-disaggregated data, SDG 4 alignment. Baseline in 3 months.
T – Task: Develop 5 SMART outcome indicators for quarterly donor reporting.
O – Output: Table with indicator statement, unit, data source, frequency, baseline/target.
F – Filter: Do-no-harm, culturally appropriate, no advanced stats, exactly 5 indicators.
Pick an evaluation task (optional)
Choose a scenario to auto-fill the Task field, or write your own:
Build your RCTOF prompt (auto-generates)
Your ready-to-copy prompt (RCTOF format):
Fill RCTOF fields above to see your ready-to-copy prompt.
Test with real AI
Copy your prompt and use ChatGPT / Claude
Paste the prompt generated above into ChatGPT (or Claude, Gemini). Evaluate the output, then refine your RCTOF components to improve results.
Open ChatGPT →Reflective learning log
After testing with ChatGPT, note what worked and how you would refine the prompt using RCTOF:
Pro tips: improve your RCTOF prompts
- Clear: Task must be specific (“list 5 indicators” not “help me with indicators”).
- Complete: Fill all 5 RCTOF components — Role, Context, Task, Output, Filter.
- Contextual: Add sector, geography, donor requirements, participant details.
- Constrained: Use Filter to set ethics, length, language, and exact counts.
- Example improvement: Instead of vague Filter, write “do-no-harm, max 200 words, avoid male-centric language”.
Now you're ready — copy the prompt, test with ChatGPT, and iterate using the RCTOF framework.
The courses and articles are developed by a team of experienced evaluators, collaborators, authors, and software developers, guided by Fation Luli. EvalCommunity Academy combines practical expertise in Monitoring & Evaluation and International Development with the latest advances in AI to create high-quality, accessible, and practical learning experiences for professionals worldwide.
