AI Usage Style Test: What the 7 Factors Actually Measure

Say "AI test" and most people picture a score. Seventy-two out of a hundred. Intermediate. Slightly behind the colleague who won't stop talking about prompts. That picture assumes everyone is on the same road, just at different mile markers, and that the point of testing is to find out how far you still have to drive.

But spend a week watching how the people around you actually work with AI and the single road disappears. One colleague invents uses nobody asked for and documents none of them. Another never opens a chatbot but catches every fabricated citation that crosses her desk. A third has quietly wired three tools together and told no one. These aren't different amounts of the same skill. They're different directions, and even big-company research is drifting the same way; Slack's Workforce Lab sketched five AI personas along similar lines.

That's the idea behind an AI usage style test. Instead of one number, RELACTIS measures seven factors: Creation, Crossover, Autonomy, Advocacy, Systemization, Verification, and Engagement. None of them is a grade. Each is a dial with useful behavior at both ends. Here's what each dial is actually listening for.

Creation: invent or refine?

Creation measures your itch to try unproven uses. If you score high, you ask "could it be used like this?" before anyone asked you to, you point features at problems the manual never mentions and treat failed experiments as tuition. It's the raw material of every genuinely new workflow, and it's the signature of The Pioneer.

A low Creation score isn't a lack of imagination. It usually means you'd rather run proven plays well than gamble on speculative ones, and someone has to. High-Creation people produce prototypes; low-Creation people are the reason anything ever reaches a dependable version. Every team that ships has both ends of this dial, whether or not anyone planned it that way.

Crossover: connect or go deep?

Crossover is about breadth, pulling an idea from one domain into another, chaining tools nobody thought to combine, finding value in the gaps between products. High scorers are collectors of parts whose favorite question is "what if we connect these?", the restless energy of The Remixer.

The low end of Crossover is depth, and depth is not a consolation prize. Pushing one tool and one purpose further than anyone around you, the pattern behind The Artisan, is what sets a team's quality ceiling. Wide-and-shallow discovers what's possible; narrow-and-deep makes it excellent. Neither can do the other's job.

Autonomy: self-serve or guided?

Autonomy measures how you prefer to get from "never touched it" to "using it daily." High scorers read the docs, poke at settings, and figure things out alone, the quiet self-drive of The Self-Learner, who has usually adopted a new feature before the training session about it gets scheduled.

Scoring low here just means you learn best through people: show me once and I'll run it correctly forever, or better yet, let me find the person who already knows. Faithful reproduction is badly underrated. It's why training works at all, and knowing who to ask is a legitimate style of getting results, not a failure to have your own. The self-taught, meanwhile, often can't explain how they learned anything, which is its own limitation.

Advocacy: broadcast or keep it quiet?

Advocacy measures whether your discoveries leave your desk. High scorers translate: "for your work, here\'s how you\'d use it." They recommend the right tool for someone else's problem, and generally treat a private win as unfinished until at least one other person has it too.

Low Advocacy is private practice, and there's nothing wrong with it. Deep skill is mostly built in silence, and not every experiment needs an audience. The two ends cover each other's blind spots: loud without practice is noise, and skilled without sharing is invisible. If you've ever discovered that the best AI user on your team is someone nobody ever hears from, you've met the low end of this dial.

Systemization: improvise or templatize?

Systemization measures the urge to turn a one-off win into something repeatable: a template, a checklist, a workflow that produces the same result no matter who runs it. High scorers say "already made it a template" and mean it. This converter instinct, embodied by The Systemizer, is one of the rarest behaviors in any organization, which is exactly why unshared inventions keep evaporating when their inventors change jobs.

Low scorers are improvisers: every task gets a fresh approach, tuned to today's specifics. That flexibility matters more than it sounds, because every system eventually meets the case it wasn't designed for, and when it does, the person who never needed the template is the one who isn't stuck.

Verification: check or trust?

Verification measures your reflex to ask "what's the source for this?" High scorers read AI output the way an editor reads a first draft: useful, promising, and not yet true. As generating plausible text gets cheaper, this checking instinct, the core of The Verifier, only becomes more valuable.

But low Verification is not carelessness; it's momentum. Low scorers are willing to work with drafts, keep speed up, and accept that a fast 80% often beats a slow 100%. If everyone verified everything, nothing would ship on time. If no one verified anything, errors would ship instead. This is the clearest case of a dial that only works as a dial, a team needs the tension between its two ends, not a winner.

Engagement: how close do you actually stand?

Engagement is the plainest of the seven: how much of your working life does AI actually touch? High scorers have it wired into daily routines and notice within hours when a model starts behaving differently. Low scorers keep their distance, and this is the factor where a "low score is fine too" reading surprises people most.

Because distance carries information. Sometimes low Engagement is a first step that hasn't happened yet, the "looks useful" of The Quiet Abstainer, often the largest silent group in a workplace. Sometimes it's deliberate skepticism that spots what enthusiasts have stopped seeing. Watching before moving is a position, not an absence, and a test that treats non-users as zeros learns nothing from the people it most needs to understand.

Seven dials, fifteen shapes

You might be wondering how to max out all seven. The assessment doesn't work that way: there is no total. The seven values come out of separate questions, and adding them together produces a number that means nothing. What you end up with is never a total. It's a shape.

That shape is how the typology works. The assessment reads your strongest signals across the seven factors, and their combination maps to one of the 15 AI usage types. High Creation with low Systemization points toward The Pioneer, whose inventions outrun his documentation. High Systemization points toward The Systemizer, who turns other people's inventions into assets. High Verification with low Engagement lands near The Critic, the skeptic whose eye for flaws is real. You also get the full seven-factor radar alongside your type, because the type is a headline, not the whole story.

The point of a usage style test, in the end, isn't to rank you. It's to name the direction you already lean, so you can stop treating it as a gap to apologize for and start using it on purpose.