How It Works

From your existing materials to a deployed AI-augmented course — six steps.

The Jobs page explains why Crucibl exists. The Methodology page explains what it is. This page is how a faculty member actually moves from a Canvas course they already teach to a Crucibl-augmented version their students engage. Faculty stays in control of pedagogy throughout. Crucibl provides the methodology, the runtime, and the audit trail — not the content.

The Six Steps

Upload, diagnose, configure, publish, choose pathway, operate.

Each step builds on the previous. Steps 1–4 happen before the term starts and produce a configured Crucibl course attached to the faculty's Canvas shell. Step 5 is operational — how students pay. Step 6 is what happens during and after the term.

A faculty member at home with a laptop, beginning the Crucibl onboarding process — uploading their existing course materials
01

Upload your existing course materials

Faculty drops their syllabus, Canvas content (modules, assignments, rubrics), and any supplementary materials into Crucibl in one motion. Crucibl reads them, parses out structure, and presents the course as it currently exists — modules, assignments, learning outcomes, assessment cadence — for the faculty member to confirm. No retyping. No data entry. The starting point is the course you already teach.

Faculty also adds optional Inspirational Principles at this step — canonical quotes, scriptures, professional codes of conduct, mission statements, or favorite-author wisdom that they want the AI agents to bridge to during student conversations. This is where the faculty's worldview and pedagogical voice become part of the course architecture, not just decoration.

02

Run the free diagnostic

Crucibl's diagnostic agent reads what was uploaded and produces a structured assessment of how AI-ready the course currently is — scored against the 10-Principle Framework. The output is a numerical score per principle, a narrative gap analysis, and a list of specific places where the course is at risk for the documented failure modes (asymptotic grading, AI-as-cheating-detection-arms-race, cognitive-load displacement, missing audit-trail integrity). The diagnostic is free, takes about 10 minutes of faculty time, and produces a downloadable PDF.

Faculty who decide not to proceed past the diagnostic still leave with something useful: a research-grounded read on where their course sits relative to the field. Faculty who do proceed use the diagnostic gaps as the starting agenda for Step 3.

A faculty member configuring a course through the Co-Design Wizard, with a laptop showing the multi-phase configuration interface
03

Configure with the Co-Design Wizard

Sixteen phases. The wizard walks the faculty member through configuring the course architecture, persona role-archetypes, scaffold maps, constraint sets, activity patterns, capstone game-dynamics, calibration anchors, and the audit-trail rubric. The wizard pre-populates a draft from what was uploaded; the faculty edits it. AI suggestions across every phase. Faculty validates each phase before moving on.

The output is a complete course specification — a JSON document that captures the faculty's pedagogical decisions in a form the runtime can execute. Faculty can stop and resume across multiple sessions; nothing is lost between visits. Configuration time is typically 10–20 hours total, spread across 2–3 weeks. White-glove faculty success support is bundled in the per-student price; concierge curriculum-design help is available as a separate engagement for institutions that want it.

A student engaging an AI mentor inside the Crucibl runtime, showing a conversational interface on a laptop
04

Publish to Crucibl

The configured course goes live. Crucibl provisions personas, sessions, scaffold maps, audit trail wiring, and the constraint sets the faculty configured. Crucibl integrates with Canvas through LTI 1.3 (open standard, no proprietary lock-in) — students enter Crucibl from inside their Canvas course shell, and grades flow back to the Canvas gradebook. Faculty does not have to retrain their students on a new login or a new system.

From this moment forward, students engaging the AI-mediated parts of the course are working inside Crucibl's runtime. Each interaction is bounded by the constraint sets, recorded in the audit trail, and graded against the rubric the faculty configured. The faculty's role shifts from "police AI use" to "design and observe AI use."

05

Choose your payment pathway

Crucibl is sold as course materials at $150/student per class (with $99–125 pilot pricing for the first ten cohorts). Title IV financial-aid eligible the same way printed textbooks are — students can pay through Pell Grants, federal student loans, and standard course-materials channels. Three pathways, depending on what fits the institution:

PATHWAY 1

Direct Stripe

Faculty adopts, students pay $150 directly through a Stripe-hosted checkout at enrollment. Lowest friction for individual faculty pilots and early adoptions. Default pathway during 2026.

PATHWAY 2

Bookstore

Crucibl listed in institutional bookstore catalogs (Follett, Barnes & Noble College, Akademos) as course material with an ISBN-equivalent identifier. Students purchase at registration; bookstore remits to Crucibl monthly minus their margin.

PATHWAY 3

Inclusive Access

Institutional bulk subscription — institution prepays student volume at a discounted rate and bills students through tuition or a course-materials fee. Strongest fit for community colleges and regional state systems where Inclusive Access is already the norm.

A faculty member reviewing the Riley dashboard, looking at structured cohort progress data on a laptop
06

Operate, observe, refine

The course runs. Students engage personas inside Crucibl, complete sessions, generate audit trails. Faculty has a dedicated visibility layer — the Riley dashboard — that surfaces cohort patterns, individual student progress, and generated weekly briefings flagging where students are stuck or where the design is producing unexpected outcomes. Faculty reviews audit trail PDFs the same way they would review traditional graded student work.

Mid-term and end-of-term, faculty refines the configuration based on what the data shows. Edits made now become better defaults for the next cohort. The configuration stickiness compounds — by the second or third cohort, the course is more tightly tuned to the faculty's specific pedagogy than any third-party textbook could be, and the audit trail provides defensible documentation at curriculum committee, accreditation review, and tenure-and-promotion files.

What Faculty Walks Out With

Four artifacts. Defensible. Portable.

At the end of the term, the faculty member has four documented artifacts that travel with them across institutions and that hold up at external review. These are the deliverables Crucibl produces — the methodology made tangible.

Course Architecture Document

A faculty-authored redesign of the course around the 10-Principle Framework — modules, assignments, scaffold map, persona territories, assessment cadence, calibration anchors. The blueprint that drove every other component, exportable as a versioned PDF.

Constraint Set Library

Per-assignment AI-use rules — what students may use AI for, what they must do without it, how the audit trail captures the boundary. Constraint is the lever, not access. Reusable across courses and editable per term.

Audit Trail Report

A 7-field log of every student-AI interaction across the term — visible, reviewable, and gradable evidence of process. Defensible at curriculum committee, accreditation review, integrity hearings, and tenure-and-promotion files.

Before-and-After Outcome Package

Pre/post outcome measurement comparing student performance with Crucibl against historical baselines or against a non-Crucibl section. The package institutions need to defend the AI-augmented design to their boards and accreditors.

A faculty member and a student sitting side-by-side reviewing a printed audit trail PDF together, the faculty member pointing at a section of the document
Faculty and student reviewing an audit trail PDF — the gradable artifact of an AI-mediated learning session, visible to both.
What This Is Not

Three things faculty should not expect.

Honest framing matters. Crucibl is not a magic AI course-builder, not a Canvas replacement, and not a black-box AI tutor. The pedagogy belongs to the faculty.

Not a magic course-builder

Crucibl does not generate a finished course from a syllabus upload. The Co-Design Wizard pre-populates a draft, but the faculty member configures, edits, and authors the pedagogical decisions. Crucibl is the methodology and runtime. Pedagogy belongs to the faculty.

Not a Canvas replacement

Crucibl runs alongside Canvas through LTI 1.3, not instead of it. Canvas is the system of record — gradebook, modules, administrative envelope. Crucibl handles the AI-mediated parts of the course. The two coexist by design.

Not a black-box AI tutor

Every persona's behavior is configured by the faculty, versioned in the database, and reviewable at any time. Every student interaction is recorded in the audit trail. Faculty can see exactly what the AI is doing, change it, or stop it. No opaque LLM weights making pedagogical decisions on their own.

A small university classroom with a professor at the front gesturing toward a projection screen while students lean forward, some typing on laptops, some taking handwritten notes — what an AI-augmented course looks like in operation
An AI-augmented course in operation — what the six steps actually deliver into the room.
Get Started

Run the free diagnostic — or talk to Chris.

The diagnostic is free, takes about 10 minutes, and produces a structured read on where your course sits relative to the 10-Principle Framework. Even if you don't move past the diagnostic, you walk out with something useful.

Get the Free Guide Talk to Chris