Teachers author the terms.
The learning goal, source material, agent role, and limits begin with the teacher.
Learning changed. Course software did not.
Students already work with AI. Teachers need a way to shape that work, see the learning inside it, and make the final judgment. Atlas is being built around that reality.
AI should change the work itself.
Putting a chatbot beside an old assignment does not make the assignment fit for the AI era. It often leaves the teacher with less visibility and the student with less reason to think.
Atlas starts with the assignment. The teacher decides where AI should help, what course material should ground the session, what the student still has to decide, and what evidence should return for review.
Assessment follows the same record. The course can then grow around it, with materials, people, grades, integrations, and agents sharing context instead of living in separate tools.
The responsibility stays human.
Atlas can make the work more legible and the course easier to operate. It should not hide who set the terms, who did the work, or who made the decision.
Students still make the call.
The agent can question and challenge. It cannot quietly turn the work into an answer generator.
The process remains visible.
Claims, sources, revisions, feedback, and final work stay connected.
Teachers decide what is final.
Suggested feedback and assessment remain reviewable before they affect the course record.
Change the assignment first.
The rest of the learning system can follow the work.
Request pilot access