Select a score 1–5 (or N/A) and add optional notes per category. Overall score calculates automatically.
Analyzing conversation…
Category scores
AI feedback
Scoring criteria
5 categories, scored 1–5. Weights total 100%.
Issue understanding
Fully understood the problem before attempting a fix.
Asked targeted probing questions to clarify ambiguityAccurately identified the root cause and customer contextRecapped the issue back to the customer to confirm understanding
25%
Troubleshooting quality
Investigated thoroughly and methodically before reaching a conclusion.
Followed logical, step-by-step troubleshootingUsed all available resources — docs, admin tools, logs, colleagues, past ticketsEscalated only when genuinely needed; avoided premature escalation
25%
Process & documentation
Followed defined support processes and left a complete paper trail.
Linear ticket is clear: steps to reproduce, expected vs actual, environment, customer impactAll relevant links included — Slack threads, Looms, related tickets, internal notesFollowed the defined escalation and documentation process
25%
Resolution & follow-through
Closed the loop with the customer clearly and delivered on commitments.
Set clear, realistic expectations for next steps and timelinesFollowed up as promised — no dropped ballsCustomer confirmed understanding and left with a clear outcome
15%
Communication style
All customer-facing messages were clear, professional, and empathetic.
Warm and empathetic — never dismissive, robotic, or impatientFree of typos, grammatical errors, and unclear phrasingTone matched the situation and the customer's emotional state
10%
Scoring rubric
Score
Label
Meaning
5
Excellent
Exceeds expectations; model example
4
Good
Meets expectations, minor room to improve
3
Acceptable
Mostly adequate but noticeable gaps
2
Needs work
Below expectations; customer experience impacted
1
Unacceptable
Critical failure; requires immediate coaching
N/A
Not applicable
Criterion not relevant to this conversation
How calibration works: Rules you add here are injected into every AI review prompt so the AI scores consistently with your team's standards. Add rules when the AI gets something wrong, or to encode edge cases you've already discussed as a team.
Add a rule
Active rules
No rules yet. Add one above.
No reviews saved yet.
Proxy configuration
Your deployed Cloudflare Worker URL. See SETUP.md for deployment instructions.
Saved in your browser.
Data
Review history and calibration rules are stored locally in your browser.