Evidence, not guesswork
How AI fluency scoring works
Five equally weighted dimensions measure demonstrated AI practice in your role. Each ranges from 1 to 5. All five at 1 means 1/10; all five at 5 means 10/10. A single strong dimension cannot stand in for the others.
The calculation
Total = 1 + (average dimension score − 1) × 9 ÷ 4
Rounded to one decimal. Five 3s give 5.5/10; five 4s give 7.8/10. Missing evidence is not a low score: the assessment stays provisional without a total until every dimension can be assessed.
From 1 to 10
These labels summarize the scale; your total comes from your actual breakdown, not a separately assigned label.
What earns each dimension score?
Tool selection and breadth
- 1/5Uses a tool without considering fit
- 2/5Chooses an approved tool with help
- 3/5Independently selects suitable tools and explains limitations
- 4/5Compares or combines appropriate tools
- 5/5Builds repeatable, well-justified tool choices others can reuse
Task complexity
- 1/5Simple isolated requests
- 2/5Routine tasks with guidance
- 3/5Completes multi-step role tasks independently
- 4/5Handles complex constraints and exceptions
- 5/5Designs robust workflows for complex, high-value work
Prompting, iteration and verification
- 1/5Accepts a first answer without checking
- 2/5Adds context and makes basic corrections
- 3/5Uses clear constraints, iteration and output checks
- 4/5Systematically evaluates alternatives and failures
- 5/5Uses reproducible evaluation, risk controls and improvement cycles
Evidence of OKR impact
- 1/5No demonstrated connection to the OKR
- 2/5Describes a plausible connection
- 3/5Shows a concrete outcome tied to a key result
- 4/5Measures a meaningful improvement against a baseline
- 5/5Demonstrates sustained, repeatable contribution with credible evidence
Consistency and transfer
- 1/5Isolated practice
- 2/5Occasional reuse
- 3/5Regular practice and lessons applied to new work
- 4/5Sustained, adaptable habits across tasks
- 5/5Sustained practice plus reusable learning that improves team capability
Fairness and improvement
You do not need lots of tools, a senior job title or polished writing to score highly. Appropriate use of your approved toolset matters. Consistency uses rolling learning history. OKR impact needs credible outcomes, not claims of business completion. Every new reflection explains the evidence, confidence and next steps; older scores are preserved without invented breakdowns.
AI assessments can be imperfect. They reflect recorded evidence, not your value as an employee. Prompt-practice scores assess individual prompts and are separate from this fluency rubric.