A research-grounded field guide for faculty, department chairs, center directors, deans, foundation funders, and grant reviewers. Drawn from the Ensign College FIN 485 deployment (case study currently in peer review), the Cognitive Load Theory foundation, and the buyer-side rejection of bundled-AI documented in April–May 2026.
A field guide for faculty teaching in the cat-and-mouse era of AI in higher education — and the institutions, funders, and grant programs supporting them.
Fill out four fields and we'll email you the PDF. You'll also receive periodic Crucibl updates — research findings, pilot announcements, launch notifications. No spam. Unsubscribe at the bottom of any message.
Detection tools have plateaued. Integrity policies create cynicism. Free AI bundled into the LMS has produced two years of empirical evidence — Microsoft Copilot enterprise deployments, Cal State's $17M ChatGPT Edu contract, Khanmigo's "non-event" admission — that capability without methodology does not change outcomes.
The way out is the calculator-and-spellchecker arc: faculty redesign assignments around the new tool, clarify permitted use, and explicitly teach prompt crafting, verification, and ethical use. That redesign is the work. Crucibl operationalizes it. This guide is the field-level explanation of what the work actually looks like.
The guide is free. The methodology is built. The buyer-side rejection of bundled-AI is now public record. What's missing — and what this document supplies — is the operational map.
If you're being judged on student outcomes and your current AI strategy is "we have Canvas's bundled AI features turned on," this guide is the design system that turns those features into evidence you can defend at curriculum review.
The accreditation question — "how is your institution addressing AI in the curriculum?" — has become unavoidable. The guide lays out what a defensible institutional answer looks like, drawn from the Wasden FIN 485 case study currently in peer review.
Crucibl's planned NSF SBIR Phase I research proposal — currently in preparation, not yet submitted — would adapt the methodology for offline-first deployment in correctional and international low-resource contexts. The guide is the methodology overview that contextualizes the planned research and the commercialization plan.
The same self-hosted Crucibl architecture serves regulated-industry on-premises deployments. The guide explains why the methodology layer matters when the AI capability is already commodity.