Your work is the curriculum.
Generic AI curricula train everyone identically and send them home to unchanged work. Personalization fixes that — but personalization is an assessment problem, not a settings toggle. How our instrument reads the way you actually work, and builds the training path from there.
Everyone gets the same module one
Picture two people enrolled in the same AI course. One runs claims operations at a regional insurer; her week is intake queues, exception spreadsheets, and a Tuesday report nobody reads until Thursday. The other is a backend engineer who already pipes model output through CI. On day one, both watch the same video about what a prompt is. One of them is bored. The other is lost by Friday, because nothing in the course touches an intake queue. The course didn't fail them equally, but it failed them both.
This is the standard failure of generic AI training, and it isn't a content problem. The material can be excellent and the failure still holds, because the failure is structural: one curriculum, every learner, in whatever order the syllabus was authored. Train everyone identically and you guarantee the training lands on work it wasn't shaped for. Then everyone goes home — to the inbox, the queue, the codebase — and the work is exactly as it was.
What sticks is what attaches to the work you already do. Everything else is trivia with a completion bar.
Generic training lands on work it wasn't shaped for — and the work stays exactly as it was.
Personalization is an assessment problem
The industry's answer is a dropdown. Pick your role, pick your level — beginner, intermediate, advanced — and the same modules get reshuffled with your first name on the dashboard. That's a settings toggle wearing personalization as a costume, and it fails for a precise reason: it asks you to self-report a capability you don't yet have the vocabulary to describe. Nobody knows they're an L1 prompter. Establishing that is the training's job, not the intake form's.
Real personalization has to start earlier, with an instrument that reads how you actually work. Not a quiz about AI — a look at the work itself: how you write instructions when something real depends on them, how you direct an agent through a task and what you check afterward, how you'd decompose the recurring job that eats your Thursdays. Three samples of practice say more than any number of claims about practice.
So the LockedIn assessment is built as the front door, not a marketing quiz. You describe your role and the task that recurs, then do three short pieces of real work: write the prompt, direct the agent, break the workflow down. The instrument reads what you did — not what you say you can do. That is what an AI capability assessment has to mean for the output to be worth building a curriculum on.
Six dimensions, honestly scored
The assessment scores six capability dimensions — each one a thing that shows up, or doesn't, in the work you submitted:
Each dimension is scored against the platform's authored rubric — the same published standard an activated cohort would use. When the scoring model is live, it reads your exercises against that rubric, and every score carries a note saying what it observed: the instruction that specified a format but never a failure case, the agent direction with no check on the output. When the model is offline, a deterministic check runs instead — and the result says so, labeled as an offline check, not dressed up as something smarter. Either way you get evidence, not adjectives.
The placement maps to the same L0–L4 ladder the whole platform runs on. Not a new scale invented for the funnel — the actual ladder, so your position means the same thing on day one as it does at a gate review.
- AI literacy & judgment — do you know what these systems are, and when to distrust them.
- Prompting & instruction design — do your instructions specify, constrain, and anticipate failure, or just hope.
- Workflow decomposition — can you take a recurring job apart into steps a machine can hold.
- Agent direction & oversight — when you hand work to an agent, do you stay its operator or become its audience.
- Verification & quality control — what you check before output touches anything real.
- Security & governance awareness — do you know what should never go in the context window.
A score without a note saying what was observed is an opinion. The instrument doesn't ship opinions.
A path with the why attached
Placement alone is a diagnosis without a prescription. So the second half of the output is a training path — five to eight modules pulled from the authored curriculum, each with the published rubric an activated cohort would use, sequenced for what the assessment saw. Not generated filler with your role pasted in. Authored modules, chosen for cause.
Every recommendation carries its evidence. If your agent exercise handed work off and never checked it, the path includes Evaluating agent behavior — and says that's why. If your prompt read like a wish rather than a specification, Prompting as specification shows up early, with the observation attached. You can disagree with the reasoning. You never have to guess at it.
That 'because' is the difference between a recommendation engine and a colleague who read your work.
You go back to your job
The design target is plain: you go back to your job doing what you already do with stronger AI operating judgment. The program is not meant to retrain you into an unrelated role or route your week through this platform forever. A claims lead might build a safer intake queue; an engineer might put review discipline around machine-written code. Same job and real work, with the change evidenced rather than assumed.
That's what personalized AI training has to mean if the words mean anything, and it's the half of AI-native training the category conversation keeps missing. An assessment that reads real practice. A placement on a ladder that doesn't move under you. A path with reasons attached. The machinery matters, but the premise matters more: your work is the only place capability can land, so your work is where the curriculum starts.
The assessment is free, and a free account saves your profile and path. Bring the task that eats your Thursdays.