What a low score looks like on fluency, delegation practice and description discipline, with one quick diagnostic each and a realistic view of how long each takes to fix.
We named the ten dimensions last time. Naming them is the easy part. The harder, more useful question for anyone running an assessment is what a low score on each one actually looks like on the ground, in a specific person, on a specific Tuesday.
This piece takes the first three, the ones every Ops team meets first because they sit closest to the surface: fluency, delegation practice, description discipline. Each gets the same treatment: what a low score actually looks like, and what fixes it.
A low fluency score rarely looks like refusal. It looks like a staff member paraphrasing their request to a colleague who "does the AI thing" rather than opening the tool themselves. It looks like someone assuming the tool remembers last week's conversation. It looks like never having opened a new conversation window on purpose, because nobody showed them where it is.
None of this is a character judgement. It is unpractised operation, and it is the cheapest of the three to fix: a single supervised afternoon, working through five real tasks from that person's own job, closes most of the gap.
A low delegation score shows up as one of two opposite failures, and both are common in the same organisation at the same time. Under-delegation: a competent user who never hands over anything above the trivial, because nobody has told them where the line sits. Over-delegation: someone handing over a judgement call that needed domain expertise, because the tool answered fluently and fluency read as competence.
The tell is the same in both directions: ask the person to name one task they would never hand over and one they always would. A fluent answer, given without hesitation, means the judgement has formed. A long pause, or an answer that lists features instead of tasks, means it has not.
A low description discipline score is visible in the prompt history, not the output. Look at someone's last five requests. "Make this better." "Fix this." "Do the report." One vague request in five is normal. More than one is the pattern: the task specification lives entirely in the requester's head, which means nobody else, human or tool, could execute it correctly without a follow-up conversation.
The cost is not just a worse first draft. It is the requester re-running the same vague prompt three times hoping for a better answer, instead of noticing that the answer will not improve until the request does.
A low score is not a character flaw. It is an unpractised skill with a visible fix.
Naming all three together invites treating them as one training problem. They are not. Fluency is genuinely cheap: an afternoon, a checklist, done. Description discipline is a habit, built over a week or two of someone correcting their own prompts against real feedback. Delegation practice is neither: it is judgement, accumulated from exposure to real tasks going right and wrong, and no single training session manufactures it. Organisations that budget the same two-hour session for all three consistently over-deliver on fluency and under-deliver on delegation, then wonder why the readiness score has not moved on the dimension that mattered most.
Fix fluency first, in a single afternoon; build description discipline over a week of deliberate practice against real feedback; expect delegation practice to take months, because it is judgement, not a skill taught in a session.
No external statistic cited; this article extends the organisation's own assessment methodology rather than presenting third-party evidence.
If this is the question on your desk, a thirty-minute call tells you whether the service fits, or that you do not need us yet.