Real AI fluency is five capabilities, from prompting to judgment to influence, measured against what each role actually needs. A one-size score tells you nothing. A role-calibrated read tells you exactly who is ready, who is stuck, and what the next step is for each person.
Why 'are we using AI?' is the wrong question
Most leaders can tell you whether their teams have access to AI tools. Very few can tell you whether people are actually fluent with them, or what fluency even means for a given role. A licence count is not adoption, and a company-wide training day is not fluency. The useful question is narrower: can this person, in this role, get reliable value from AI in their real work?
That question cannot be answered with a generic readiness quiz, because fluency is not a single skill. It is a set of capabilities that show up differently depending on what someone is hired to do.
The five dimensions of AI fluency
Drawing on international AI literacy frameworks, including work from UNESCO, the US Department of Labor and Anthropic's AI fluency research, fluency resolves into five measurable dimensions grouped by three questions.
Foundation: can they talk to AI and understand what it does?
- Prompting & direction, writing clear instructions and delegating a task so it actually lands.
- AI & tool literacy, knowing which tools, how deeply, and where their capability ends.
Application: can they use AI productively and safely?
- Workflow integration, moving from isolated dabbling to AI embedded in the daily job.
- Critical thinking & judgment, verifying output and catching hallucinations instead of pasting.
Amplification: do they multiply AI impact beyond themselves?
- Growth & team influence, learning agility, sharing what works and championing adoption.
Why role calibration is everything
A one-size rubric scores a salesperson and a support agent on the same scale, which tells you nothing useful. The bar for 'good' is different for each, because the work is different. A senior consultant who cannot delegate research to AI has a bigger gap than a junior who has not yet automated a weekly report.
That is why fluency should be measured against a profile: the person's company function, their role, their day-to-day work and their seniority. Score everyone against their own profile's benchmark, and the gap you see is the gap that actually matters for that role, not an artefact of a generic test.
The point of the score is not a grade. It is knowing the exact next rung for each person, so training targets the real gap instead of the whole company at once.
From a score to a plan
Each dimension lands on a five-level maturity ladder, from absent to multiplier: absent, experimental, functional, integrated, and multiplier. Knowing where someone sits, and where their role needs them to be, turns a vague upskilling budget into a precise sequence of moves.
This is what our AI Litmus module measures. Instead of a multiple-choice quiz, each person has a short, private conversation. We watch how they actually work with AI across four moves, delegation, description, discernment and diligence, then score it with our own engine and calibrate the result to their role. You walk away with recovered hours, an ROI case you can defend, and the exact first step for every person.
See this on your own teams.
A private walkthrough, calibrated to your roles. About two weeks.
Frequently asked
What is AI fluency?
AI fluency is the ability to get reliable value from AI in real work. It spans five dimensions: prompting and direction, AI and tool literacy, workflow integration, critical thinking and judgment, and growth and team influence. It is distinct from simply having access to AI tools.
How do you measure AI fluency in employees?
The most accurate method is a short, structured conversation that observes how a person actually works with AI, rather than a multiple-choice quiz that tests recall. Responses are scored across the five fluency dimensions and calibrated to the person's role, so the result reflects what their job actually requires.
Why not just use a generic AI readiness test?
Generic tests score every role on the same scale, so a salesperson and a support agent are judged identically even though their AI needs differ completely. A role-calibrated assessment measures each person against their own profile's benchmark, which makes the gaps and the training priorities meaningful.
What are the levels of AI maturity?
A common five-level ladder runs from absent (AI is not part of the work), through experimental, functional and integrated, to multiplier (the person drives adoption and builds for the team). Each fluency dimension can be placed on this ladder to show the next step.
Shobhit Khandelwal is the founder of VMS Culture Labs, on a mission to measure what most leaders only guess at: how fluently their teams truly work with AI, and the hidden cost of how people behave at work. He is out to replace workplace guesswork with evidence, and build the kind of workplaces the next generation deserves.
Connect on LinkedInOccasional, high-signal notes on measuring AI fluency and culture cost. No spam.