Write & Improve vs BAND9AI IELTS Writing
CEFR practice · IELTS rubrics · May 2026
Direct answer
Write & Improve builds general English writing with CEFR-level feedback from Cambridge; BAND9AI scores IELTS Writing against Task Response, Coherence, Lexical Resource, and Grammar on exam-style prompts. Write & Improve is excellent for writing habit and broad competence—it is not an IELTS band oracle. BAND9AI is built for students who need to know which IELTS descriptor caps their Task 2 at Band 6. Use both if you want volume plus rubric diagnosis.
Two different scoring models
Write & Improve CEFR levels (A1–C2) on varied Cambridge tasks
BAND9AI IELTS bands on Task 1/2 with four criteria split
Gap CEFR labels do not map 1:1 to IELTS Task Response rules
Side-by-side comparison
| Dimension | Write & Improve | BAND9AI |
|---|---|---|
| Task types | General English prompts | Academic/GT IELTS Task 1 & 2 |
| Feedback frame | CEFR competency levels | IELTS public band descriptors |
| Timed mode | Optional; not exam-default | Exam-style timing available |
| Best for | Writing volume and confidence | Mid-band plateau and TR leaks |
Combined workflow
- Write & Improve: 2–3 short tasks weekly for fluency and error exposure.
- BAND9AI: one blind IELTS Task 2 every 7–10 days with criterion scores.
- Compare whether CEFR praise hides thin Task Response on IELTS prompts.
- Validate with calibration anchors before booking.
Key takeaways
- Write & Improve = general writing gym; BAND9AI = IELTS rubric diagnosis.
- Do not convert CEFR levels directly to IELTS bands.
- IELTS Task Response rules are stricter than general essay tasks.
- Pair volume tools with criterion-level IELTS checks.
FAQ
Useful for writing habit—not for IELTS Task Response scoring or timed exam simulation.
Rough overlap only; IELTS uses task-specific descriptors and penalty rules CEFR does not replicate.
Yes: Write & Improve for volume; BAND9AI for IELTS rubric diagnosis on blind prompts.
Know your IELTS Writing leak—not just your CEFR label.
Get Writing Reality Check →