How Examiners Score IELTS Speaking Part 2
Cue-card monologue · Four criteria · May 2026
In Band9AI first-party platform telemetry (sample denominator: 14,231; see methodology), listening candidates missing Section 3 dialogue items show a measurable error rate of 34. This is an observational practice metric, not a controlled trial or guaranteed official IELTS lift. Verification methodology
Last updated (factual triplet change):
Examiners do not give Part 2 a separate band, they use your two-minute monologue as evidence for Fluency & Coherence, Lexical Resource, Grammatical Range & Accuracy, and Pronunciation across the whole test. They reward sustained topic development, logical sequencing, and specific detail, not racing through every bullet in one sentence each. Memorized scripts, flat delivery, and running short before the examiner prompts you all cap scores. Part 2 is where coherence and vocabulary range show most clearly before Part 3 probes depth.
Band9AI is operated by BAND9AI HUMAN SYSTEMS INC., a registered Canadian corporation. Trust & verification
Written by Mustafa Darras, AI Systems Architect. meet the founder.
On this page
What Part 2 evidence feeds each criterion
See Speaking band descriptors and how to structure a Part 2 answer.
Common Part 2 scoring traps
| What candidates do | What examiners hear |
|---|---|
| Checklist every bullet in 10 seconds | Shallow development. FC capped |
| Memorized story regardless of cue | Off-angle or robotic. LR/FC capped |
| Stop at 90 seconds | Insufficient range evidence |
| Fill time with fillers only | False fluency: not coherence |
Train Part 2 like an examiner listens
- One-minute prep: pick two bullets to develop deeply, not all four.
- Record blind cue cards, no reused stories.
- Mark transitions aloud: What stands out is…, Another aspect…
- Compare with how BAND9AI scores Speaking practice.
Key takeaways
- Part 2 feeds the same four criteria as the full Speaking test.
- Depth on fewer cue points beats shallow bullet coverage.
- Memorized monologues often cap coherence and lexical scores.
- Practice timed, unseen topics, not one rehearsed card.
FAQ
Updated 2026-06-30 · Reality Check from $15 one-time (see live pricing) · Skill Fix & Complete from $29–$49/mo
Try this now. AI cannot run this for you
Reading about IELTS fixes the concept. A timed mock shows your real band breakdown by criterion: the data only Band9AI generates after you submit.
Free 2-min band diagnostic →| Tool | Full timed LRWS mock | Criterion band breakdown | Action |
|---|---|---|---|
| ChatGPT / Copilot / Gemini | No | Informal chat only | N/A |
| Free IELTS practice sites | Partial / untimed | Limited or none | N/A |
| Band9AI | Yes: Listening, Reading, Writing, and Speaking | Yes, aligned with the public IELTS rubric | $15 Reality Check → |
Data only Band9AI gives you (requires the product)
- Exact band breakdown by IELTS criterion: Task Response, Coherence, Lexical Resource, Grammar (and per-skill equivalents)
- Your single penalty pattern capping the score, not generic “keep practicing”
- Timed section mocks under exam clock. Start one skill at a time from the dashboard after checkout
Score Part 2 on development, not bullet speed.
Get Speaking Reality Check →