Introducing Casiana

Casiana is a K-12 learning management system for schools that teach through projects. It is built by Casiana AI LLC, and this is the first thing we have written publicly about it.

The short version

Most learning management systems store one grade per assignment per student. That is a reasonable model for a course made of quizzes and problem sets. It is the wrong model for a course made of projects, because a project is deliberately integrative: one deliverable is real evidence about a student's thinking, their writing, their speaking, their collaboration, and their self-direction, all at once.

Casiana stores those separately. A teacher tags a project with the learning outcomes it assesses, the student submits one piece of work, and the teacher scores each outcome against its own rubric. Students see per-outcome scores and per-outcome trends. Course grades aggregate by outcome, not by assignment average. An administrator can ask how a student is doing on written communication across every class they take, and get an answer.

We call that multi-outcome scoring, and it is the reason the platform exists as something separate rather than as a plugin.

Where AI fits, and where it does not

Casiana is AI-native in a specific and deliberately limited sense.

Students get an AI coach that reads a draft and gives formative feedback aligned to the outcomes their teacher actually selected for that project — not generic writing advice, but "your evidence does not yet connect to your claim," against the same rubric the teacher will use. It runs before submission, when feedback can still change the work.

Teachers get drafting help: rubric generation, activity scaffolding, driving question suggestions, and pre-filled rubric levels to review.

What AI does not do is assign a grade. Every AI suggestion is labeled, every one is editable, and a human teacher is the final assessor of record. Submission content is stripped of personally identifying information before it goes to a model, AI interactions involving student data are written to the audit log, and the whole capability is opt-in. Those are not marketing positions; they are what FERPA compliance requires of anyone doing this seriously.

It does not need to replace what you already use

A school that runs on Google Classroom does not have to abandon it to use Casiana. The multi-outcome gradebook is useful on its own, on the projects that matter, alongside whatever handles day-to-day assignment distribution. We would rather be additive and get used than demand a rip-and-replace migration in July and get shelved.

The frameworks underneath

The default rubrics that ship with Casiana come from public frameworks: claim–evidence–reasoning for scientific explanation, the AAC&U VALUE rubrics for written and oral communication and for collaboration, the three-dimensional framework behind the Next Generation Science Standards, Bloom's revised taxonomy for the vertical structure of levels, and the ISTE standards for digital work. They are defaults, not doctrine — every outcome and rubric is renameable, reweightable, and replaceable per school.

Shipping defaults matters because a school evaluating a platform in October should be able to grade a real project that week, not after finishing a curriculum mapping project.

Who this is for

Project-based schools, microschools, and small networks — the places where the pedagogy is the point and the gradebook keeps getting in the way. If that is you, join the waitlist. We are looking for a small number of pilot schools who want to use this on real projects with real students and tell us where it falls short.