Crucibl — Canvas Runs Your Course. Crucibl Designs and Deploys It.

The Problem Higher Ed Is Already Living

Students use AI. Faculty police. Nobody learns.

Today, students use AI to assist with assignments. Faculty respond with detection tools, integrity policies, and assessment redesigns. The detection tools have plateaued. The integrity policies create cynicism. The redesigns are partial. Both sides escalate. Faculty spend their hours policing rather than teaching. Students spend theirs hiding what they're already doing.

This is the cat-and-mouse phase of AI in higher education. It is unsustainable. And it has produced the empirical record below.

92%

Students using AI for academic work

92% of students report using generative AI for their coursework — up from 66% one year earlier. Use is now ubiquitous across disciplines and demographics.

HEPI Student Generative AI Survey, 2025

17%

Performance drop on independent work

Students using unconstrained GPT-4 scored 17% worse than the no-AI control group on independent assessment once AI was removed.

Bastani et al., RCT, N≈1,000

2.85×

Differential harm to novice learners

Unstructured AI interaction harms novice learners 2.85× more than advanced learners — the population schools are most responsible to protect.

Lepine et al., 2026

The Microsoft Copilot Lesson

Generic capability without method does not change outcomes.

Microsoft bundled Copilot into 365 at $30 per user per month for two years. The empirical answer is in the record: Gartner ROI studies show enterprise customers cannot justify the spend. Forrester named "capability without workflow" the top failure mode. Microsoft itself has publicly experimented with non-OpenAI models because the current state isn't working.

Free AI bundled into your LMS will follow the same arc. Faculty bolting Canvas's bundled AI onto unchanged course designs replicate the Copilot experience: AI features turned on, outcomes unchanged, accreditors unimpressed. Crucibl is what method on top of that capability looks like.

What Buyers Just Decided

April–May 2026 is the buyer-side rejection window.

Five public events, in a single fortnight, mark the structural shift from "deploy bundled AI and see what happens" to "demand methodology and outcome evidence." These are decisions buyers have already made — not Crucibl forecasts.

Cal State faculty petition · 3,300+ signatures

California State University's $17M ChatGPT Edu contract expires June 30, 2026. Faculty across the 23-campus system petitioned the Chancellor not to renew, calling it "a general-purpose chatbot that is not designed, trained, or optimized for education."

California Faculty Association, May 2026

Harvard FAS multi-vendor pivot

Harvard's Faculty of Arts and Sciences announced it is phasing out ChatGPT Edu and expanding access to Anthropic's Claude — explicitly framed as ongoing, multi-vendor governance. The most prestigious U.S. university publicly refused single-vendor commitment.

Harvard Crimson, April 28, 2026

ASU "Atomic" walk-back

Arizona State soft-launched a consumer-facing AI course-builder that scraped faculty content without consent. President Michael Crow walked it back: "wasn't really ready for prime time...hadn't been evaluated yet." The exact failure mode Crucibl is built to prevent.

Inside Higher Ed, April 28–29, 2026

Sal Khan on Khanmigo

The founder of the most-watched AI-tutor product publicly admitted: "For many students it was a non-event." The strongest possible refutation of "deploy AI tutoring and see results" — delivered by the most-credible AI-education evangelist himself.

Sal Khan, Chalkbeat, April 9, 2026

When buyers across faculty, institutional administration, and the most-credible operators in the category all reach the same conclusion in a single window, the question stops being "is methodology over capability the right thesis?" — and becomes "what does the methodology layer look like, operationalized?"

The Way Out

Reframing AI as a literate, bounded tool.

Calculators were once banned in math classrooms. Spellcheckers were once derided as crutches that would destroy student writing. Both were eventually normalized — not by faculty surrendering on rigor, but by faculty redesigning what they were teaching and how they were assessing it.

The AI arc is the same, accelerated. The work is concrete: redesign assignments, clarify permitted use, explicitly teach prompt crafting, verification, and ethical use. That redesign is the work. Crucibl operationalizes the work.

"The trend this year should be toward reframing AI as a literate, bounded tool — similar to how calculators and spellcheckers were eventually normalized — by redesigning assignments, clarifying permitted use, and explicitly teaching prompt crafting, verification, and ethical use. Strategically, institutions should expect to invest in faculty and staff development so AI augments work rather than simply adding a new compliance burden."

— Nick Swayne, President, North Idaho College · Tech Outlook 2026, Campus Technology

This is what Crucibl does, in the words of a sitting college president who has no relationship to the company. Crucibl was designed to operationalize exactly this work — before Swayne wrote his quote.

How Crucibl Operationalizes the Work

Methodology first. Software second.

A faculty member and a student in a faculty office, mid-conversation — the mentorship moment Crucibl is built to scaffold

The mentorship moment Crucibl is built to scaffold — not replace.

The 10-Principle Framework is the product. The multi-agent software stack — Builder, Socratic Tutor, Critic-Coach, and Instructor Insight Agent — is the runtime that operationalizes it. Thirteen provisional patent applications pending cover the core architecture.

Course Architecture Document

A faculty-authored redesign of the course around the 10-Principle Framework. The blueprint that drives every other component.

Constraint Set Library

Per-assignment AI rules that enforce when, how, and to what extent students may engage AI. Constraint is the lever, not access.

Multi-Agent Pedagogy Stack

Builder, Socratic Tutor, Critic-Coach, and Instructor Insight Agent — each with defined roles, limits, and orchestration logic.

Audit Trail Report

A 7-field log of every student-AI interaction. Visible, reviewable, and gradable evidence of process — defensible at accreditation and integrity review.

Read the full methodology

Who We Serve

Four populations. One methodology.

Crucibl's framework generalizes across higher-education contexts where AI capability is rising faster than the design discipline to use it well — from US correctional postsecondary education to international low-resource higher ed to mainstream academic institutions to professional and corporate training.

Prisons

~30,000+ Pell-eligible postsecondary students in US correctional facilities, growing 20–40% annually post the 2023 Pell Grant restoration. Self-hosted, offline-first deployment is the focus of Crucibl's planned NSF SBIR Phase I research proposal — currently in preparation, partnership discussions with Ensign College underway.

International

Universities and continuing-education programs across Sub-Saharan Africa, South Asia, and Latin America face the AI-pedagogy gap with severe additional constraints — bandwidth, mobile-first device profiles, and language. $1B–$10B addressable market.

Academic

Faculty teaching career-motivated adult learners at community colleges, teaching colleges, regional universities, and non-elite MBA programs. The buyer-side rejection of bundled-AI happened here first — and the methodology demand is highest.

Professional Education

Financial services, defense, healthcare, Big Four professional services, and continuing professional education increasingly require on-premises AI deployment. The same self-hosted Crucibl architecture serves both correctional and corporate contexts.

See market detail

Chris Wasden, EdD

Founder · Crucibl Methodology, LLC

Past experience includes Chief Strategy Officer roles in digital health, Global Healthcare Innovation Leader at PwC (New York City), executive director of the Sorenson Center for Discovery & Innovation at the University of Utah, and M&A practice at JPMorgan. EdD in Human and Organizational Learning from George Washington University, with a dissertation on how tension creates energy in Complex Adaptive Systems.

Currently adjunct professor of finance at Ensign College, where two courses now run on the Crucibl methodology: FIN 485 (the finance capstone, with the Winter 2026 deployment that produced the case study currently in peer review, Wasden forthcoming 2026) and FIN 345 (Financial Institutions, launched May 5, 2026 as Crucibl's second production course). Author of Tension: The Energy of Innovation (2014) and Solving for Why (2018), with a three-volume book trilogy in development.

Read the full bio →

Canvas runs your course. Crucibl designs and deploys it.