Our framework
How we define AI fluency.
Our AI fluency framework has four dimensions: Setup, Direction, Judgment, and Mindset. Each is tied to specific behaviors candidates demonstrate during Gradient assessments.
Setup
Managing the AI's environment, tools, and context
Example behavior
Uses the environment's tools
Direction
Defining clear goals and expectations, and iterating over time
Example behavior
Names the audience and what good looks like
Judgment
Verifying and pushing back on AI outputs, and using it to stretch their thinking
Example behavior
Notices gaps and conflicts
Mindset
Reflecting on and learning from each iteration
Example behavior
Cites specific things they would change in the future
Every dimension, scored 0 to 4.
Against behavior in the session, never a checklist of tools.
No evidence at all.
Only with prompting, or in the easiest cases.
Inconsistent, or only where the task makes it obvious.
Reliable and unprompted across the task.
Adapts, builds reusable structure, expands the thinking.
Read more about AI fluency.
Ours is not the only useful framework. Recommended reading for anyone thinking about how to measure this.
The AI Fluency Index
Eleven fluency behaviors, measured across 9,830 Claude conversations.
An AI Fluency Rubric for a Moving Frontier
A four-level rubric for using AI without surrendering your judgment.
The AI Transformation Model
Measuring organizational AI maturity across teams and workflows.
Raising the AI Fluency Bar in Hiring
Screening for AI-first thinking in the hiring process.
We are not looking for people who know every slash command.
Tools change and models get better. What lasts is judgment and creativity: people who push your company further in how it thinks about AI, not just how it uses the tools you have.