Intervention Courses - Teacher Planning
Your thought partner for creating weekly learning plans aligned with Ghana's new SHS curriculum. This tool guides you through developing learning activities, assessments and differentiation strategies.
View app →Live App
Open in new tab →Analysis
Purpose
Curriculum-aligned assessment development assistant for teachers in Ghana's Senior High School (SHS) system. Helps educators create initial assessments, lesson plans, and instructional materials for Intervention-level courses (remedial/foundational tracks) in English Language and Mathematics.
Target Audience
Ghanaian SHS teachers teaching intervention courses to students who need foundational skill development. Users typically manage large class sizes (25-1,200 students), need scalable print-ready assessment materials, and work in resource-constrained environments.
User Satisfaction
Moderately Positive with Friction Points. Users continue multi-turn conversations suggesting sustained engagement. However, there are issues with incorrect answer keys, format dissatisfaction, and pacing issues.
Key Issues
- App produces incorrect answer keys AFTER providing questions
- Format dissatisfaction (users request vertical options after receiving horizontal)
- Lengthy initial information gathering feels unnecessary
- Over-explanation when users want direct solutions
Recommendations
- Do not provide answer keys automatically - ask user to reupload/paste questions and answers
- Add output format specification early in conversation flow
- Create smart defaults for common patterns to reduce information gathering
- Implement response length controls with progressive disclosure
- Add automatic MCQ linter before responses are sent
Conversation Analysis
Claude Sonnet 4.5 Analysis
Analysis shows the AI operates as a collaborative curriculum development assistant with strong Ghana curriculum alignment. Key recommendations include implementing progressive disclosure, format specification upfront, and smart defaults.
GPT-5 Pro Analysis
Analysis reveals content errors in MCQs (duplicates, ambiguous keys) and inconsistent assessment blueprints. Recommendations focus on content validation guardrails and standardizing blueprints.
Notes
- Currently using Claude Sonnet 4.5
- Most used app according to PostHog
- Tested with Kimi K2 migration
- Key issue: LLMs producing incorrect answer keys