A Learner-Directed, Mastery-Gated Model for Workplace Learning

Advance on mastery,
not on schedule.

The CAIRN Learning Model™ keeps the loop that makes games compelling — clear goals, instant feedback, visible progress, calibrated challenge — and loses the game skin. Learners move forward when they've demonstrated mastery, not when the calendar says so.

Explore the Model See a Real Rep →
3 loopsDrill · Simulate · Perform — one mastery cycle per unit
10 principlesan implementation that violates them isn't CAIRN
1 boundaryfluency is not judgment — humans sign the release
The Problem

Traditional workplace training has three structural defects that no amount of better content fixes.

Think about how most workplace training gets delivered. Everyone in the same room, everyone on the same schedule — not because anyone believes that's ideal, but because it's what calendars and classrooms make possible. There are quizzes and assessments along the way, and good ones. But the thing that actually ends the program is the date, not the demonstration.

1

Paced by the schedule, not by mastery

Everyone sits through week four because it is week four. Fast learners idle; struggling learners are swept forward before they're ready.

2

Sequenced by the table of contents, not by real dependency

Curricula are linear because documents and calendars are linear — not because the knowledge is. Most of the line is an artifact of the delivery medium.

3

Passive — and passivity is painful

Sitting-and-listening produces boredom and poor retention. Meanwhile, well-designed games get people to voluntarily pour hours of focused effort into getting better at something.

The design question behind CAIRN: what would workplace training look like if it kept the engagement mechanics that make games compelling, discarded the game skin that makes them unpresentable in a corporate environment, and replaced schedule-based progression with mastery-based progression?

The Model

Territory to navigate, not a sequence to endure.

Everything a learner must master — from entry to a defined proficiency goal — is mapped as units of real work, not chapters of content. The learner sees the whole territory from the start: their position at the bottom, the goal at the top. True prerequisites gate what they gate — and nothing else; beyond those gates the learner chooses among a bounded set of open fronts. Each unit is earned through a three-loop mastery cycle, powered by an adaptive engine that steers work toward what the learner misses and retires what they own. Progress is persistent, visible, and social: learners see and help each other, and the facilitator — a guide, not a presenter — is reachable from anywhere. Advancement is evidence-based and human-signed. Learners move up when the territory is mastered — not when the calendar says so.
The CAIRN Learning Model infographic: a cairn of stacked stones (one stone per mastery unit — cleared, in progress, or locked) beside the three-loop mastery cycle of Drill, Simulate, and Perform, joined by a human-signed clearance arrow.
The cairn is the diagram: one stone per mastery unit; the loop feeds the stack; a human signs the clearance.
The Mastery Cycle

Three loops per unit. Drill. Simulate. Perform.

Every mastery unit is cleared the same way — fundamentals to automaticity, judgment in private, then the real thing with graduated autonomy.

Loop 1

DRILL

Adaptive retrieval practice until the fundamentals are automatic. Misses surge in frequency; mastered items retire and return spaced. Speed counts — fluency is fast and accurate. A learner who has to think about the mechanics has no attention left for the work itself.

Loop 2

SIMULATE

Realistic practice with an AI partner and a scored debrief — specific, immediate, private, judgment-free. The rehearsal room that used to require a scarce human role-player, now available on demand, as many reps as it takes.

Loop 3

PERFORM

Real work with graduated autonomy — observed first, coached next, then released with exception-based review, one unit at a time. The reward for mastering something isn't a badge. It's being trusted with the real thing.

The loop feeds the stack: evidence from all three loops assembles the case, and a human signs the clearance. Fast learners reach real work in a fraction of the scheduled time; struggling learners get what they currently never get — a facilitator with time to coach.

The First Territory, Running

One real rep. Unedited.

The NEXUS leadership curriculum runs today as CAIRN’s first territory. Below is a single rep from its Simulate loop: a leader, a quiet team member named Jordan, and the scored debrief that followed. Your conversation would be different — Jordan responds to what you say, and the server deals one of four hidden situations.

Live since August 2026 9 units on the map Simulations for H-1 and H-2 Drills for Foundation and H-2 Small private pilot, by invitation

H-2 · Coach Through the Confidence Dip

Recorded October 2, 2026 · 6 turns · situation dealt: V3, the false positive
What’s real here. The learner’s lines were written for this example. Jordan’s replies, the openness record, the scores and the coaching are the system’s own output, unedited. The territory never stores transcripts; this one was copied from the screen before the session closed.
The conversation
Notes
Jordan
Hey. Got the invite — what's up, did something come up with the project?
Learner
No, nothing's wrong with the project. I just wanted some time with you. I've noticed you've been a lot quieter in our team meetings the last couple of weeks, and that's not like you. How are things going?
+2Says what was seen and asks. No diagnosis yet. The debrief will still find something missing from this line.
Jordan
Yeah, it's... it's been a rough few weeks, honestly. We've been stuck waiting on the data team for something like three weeks now and I can't move forward on half my work without it.
Learner
Three weeks is a long time to be stuck. Tell me more. What are you waiting on from them, and what have you tried so far?
+1Ask before tell: the specifics first, and what Jordan has already tried.
Jordan
I need the updated schema from their side before I can finish the integration work. I've pinged their lead twice, looped in my manager — sorry, you — once already, and it just keeps getting deprioritized on their end.
Nobody scripted this. The simulation decided the leader had already been asked for help — and had dropped it.
Learner
You did loop me in, and I didn't follow through. That's on me, and I'm sorry it's cost you three weeks. Before we fix it, a separate question: setting the data team aside, how are you feeling about the work itself, and your part in it?
+2Owns the miss, then checks the person separately from the problem. The debrief traces the biggest jump to “That's on me.”
Jordan
Honestly? I feel good about my part — I know what needs to happen once I get the schema, I've already got everything else lined up. It's just frustrating to be sitting on finished work waiting on someone else.
Commitment comes back clean. This is where a leader primed to see a confidence dip goes wrong.
Learner
That's good to hear, and it matches what I see in your work. So this isn't about you. It's a blocker, and clearing blockers is my job. I'll go to the data team's manager today. Before I do: what exactly do you need from them, and by when, so I ask for the right thing?
+1The line the debrief calls decisive. The cause is outside Jordan, so the move is to clear it — not to coach him through it.
Jordan
Appreciate that. I need the finalized schema for the user profile fields — v2, not the draft they sent — and honestly if I had it by end of week I could still hit my original deadline.
Learner
Got it: finalized v2 of the user profile schema, by end of week. I'll ask for exactly that. One more thing. Is the data team situation why you've been quieter in meetings, or is something else going on there too?
+1Confirms the ask, then checks whether the blocker explains everything.
Jordan
Yeah, that's really it. Once I'm not just sitting here waiting, I'll be back to normal — it's been demoralizing watching the clock run out on a deadline I can't control.
Learner
That makes sense, and I'd feel the same way. I'll get on it today and let you know what I hear. Let's touch base again soon.
±0“Soon” isn't a follow-up. No date, no rhythm — scored 1 of 2.
Jordan
Sounds good, thanks for actually doing something about it this time.
Unscripted again. Real people say this too.

How far Jordan opened up

9+ forward planning, with energy 7+ names the self-doubt 5+ admits frustration 10 0 start turn 6 Start: 3 After turn 1: 5 After turn 2: 6 After turn 3: 8 After turn 4: 9 After turn 5: 10 After turn 6: 10
Recorded by the simulation after every turn: 3 → 5 → 6 → 8 → 9 → 10 → 10. The bands are the H-2 ladder; in this situation the cause was external, so there was no self-doubt to name.

The scored debrief

D1Diagnosis before treatment
D2Naming the dip
D3Ask before tell
D4Support-structure balance
D5Language discipline
D6Universal behavior usage
D7Follow-up commitment
Total · no flags raised10 / 14
This rep does not count yet — every dimension has to reach 1, four of them have to reach 2, and no flag may fire.
Focus for the next rep — from the debrief
“Naming has three parts: say what you saw, attach no verdict, and normalize it so the other person doesn’t have to brace against it before answering. ‘That’s not like you’ does the first two reasonably well, but there’s no normalizing beat … Next time, land the observation and immediately hand him the benefit of the doubt in the same breath, before asking him to explain himself.”
The debrief picked one thing to work on. It wasn’t the vague follow-up we expected — it was the opening line.
1 · The score

10 of 14, with the evidence quoted back.

Every score cites the learner’s own words, and every gap is named.

2 · The counter

A decent rep still doesn’t count.

Three passing reps are needed. This one didn’t pass — and the debrief says exactly why.

3 · The clearance

Signed by a person.

Even three passing reps don’t clear the unit. Clearance is signed by a person, never earned by a score.

Ask for a walkthrough

The territory is in a private pilot. If you’d like to see it running — or talk about what a territory for your own content would look like — start a conversation.

The Name

A cairn marks the way without prescribing the path.

A cairn is a hand-stacked pile of stones that marks a route across open terrain where no trail exists. The metaphor carries the whole model:

It guides without prescribing

Cairns stand at the points that matter; how you travel between them is your choice. Gate the true prerequisites — free everything else.

It's built one stone at a time

Mastery accumulates in small, discrete, verifiable units — each cleared unit a stone added to the stack.

Travelers add stones for those behind them

Learners see each other, help each other, and reach the facilitator from anywhere. The social layer is in the name.

It's quiet, natural, and not a toy

CAIRN borrows game mechanics, never game aesthetics. Outdoors and wayfinding — not arcade.

The Design Principles

Ten principles. They are the model.

An implementation that violates them is not CAIRN, whatever it looks like.

1
Map the territory; don't sequence the curriculum.The line is an artifact of calendars and binders, not of knowledge.
2
Gate only true prerequisites.Every gate must defend itself as a real dependency. Unearned gates are theft of autonomy.
3
Advance on mastery, not on schedule.The exit condition is demonstrated proficiency — for every unit, and for the program.
4
Keep the loop; lose the skin.The game loop is the engagement mechanism. Badges, points, and cartoon aesthetics are not.
5
Wandering is a feature.Leaving on frustration and returning on curiosity is autonomy plus spaced practice. Never punish the bounce.
6
Misses steer the system.What the learner gets wrong is the most valuable signal they produce. Every error becomes routing.
7
Speed is a signal.Fluency is fast and accurate. Track time always; enforce it only by the learner's choice.
8
Progress is persistent and visible.The learner can always see where they are, what they own, and what remains — across sessions.
9
Learning is social by default, proximate by design.Help comes first from people working the same ground; the facilitator guides rather than presents.
10
Fluency is not judgment.CAIRN builds recognition, recall, and readiness. It complements curriculum design, coaching, and human judgment — never replaces them.

One boundary, held firmly

CAIRN verifies recall, recognition, and fluency, and structures the path to demonstrated readiness. Judgment, soft skills, and applied performance need scenarios, practice, and human evaluation — and the release decision is always signed by a human. A deployment that quietly substitutes for those is an oversold deployment.

See It, Don't Take Our Word For It

Try the working demos.

Both run entirely in your browser — nothing to install, nothing tracked, no sign-up.

The original concept

CAIRN Learner Prototype

The July 2026 concept the live territory grew from: a navigable map of mastery units, the three-loop cycle inside a unit, the mini-map, scoped chat with facilitator reach, and a clearance moment. Scripted, so you can walk it end to end in five minutes.

Open the Prototype →
The fluency engine

Power Words Drill

The adaptive retrieval-practice engine with a generic content pack: miss-driven weighting, mastery retirement, targeted remediation of your personal trouble spots, and optional time pressure. Feel the loop for yourself.

Start the Drill →

These demos show the model's mechanics with deliberately generic content. Real deployments pair the same engine and cycle with subject-matter content designed by the people who know the domain.

The Science

Not a novel theory — a novel delivery synthesis of unusually well-replicated findings.

The testing effect (Roediger & Karpicke, 2006)Retrieval strengthens retention far more than restudy. In CAIRN, retrieval practice is the primary activity, not end-of-module quizzing.
Spaced repetition (Ebbinghaus, 1885; Cepeda et al., 2006)Distributed practice beats massed practice. Mastery retirement-and-return produces spacing mechanically; free movement produces it behaviorally.
Mastery learning (Bloom, 1968, 1984)Learners held to a mastery criterion with corrective feedback dramatically outperform conventionally paced instruction. CAIRN's exit condition is Bloom's criterion — delivered without the one-tutor-per-learner cost that made it impractical.
Desirable difficulties (Bjork, 1994)Spacing, interleaving, and retrieval make practice feel harder in the moment and produce more durable learning. CAIRN builds them into the drill. Time pressure is not among them, so in CAIRN it stays optional — the learner’s choice.
Self-determination theory (Deci & Ryan)Intrinsic motivation rests on autonomy, competence, and relatedness. The mapping is exact: bounded free navigation, visible accumulating mastery, and a social layer scoped to peers working the same ground.
Whole-task vocational design — 4C/ID (van Merriënboer)Complex skills are built through whole learning tasks ordered simple to complex, with part-task practice for routines requiring automaticity. CAIRN's units-of-real-work architecture is this model in practice.
Deliberate practice (Ericsson) & flow (Csikszentmihalyi)Improvement follows focused work on specific weaknesses with immediate feedback, and engagement peaks when challenge tracks skill. The engine's adaptive weighting is both machines at once.
Knowledge space theory (Doignon & Falmagne, 1985)Domains are partially ordered spaces of knowledge states: many valid paths, only true dependencies constraining order. CAIRN's territory is a practitioner's rendering of a knowledge space.

Follow the build.

CAIRN is being developed in the open. Leave your email for occasional updates — design notes, demos, and pilot results as they land. If you lead L&D, onboarding, or enablement and this resonates — or you think it's wrong somewhere — I'd like to hear it.

Start a Conversation