AI and data
How Ensemble uses AI
Last updated: September 30, 2026
Ensemble uses AI for one job: reading the service information out of IEP documents so district staff don't have to re-type it.
What AI does today. When a district provides IEP documents or service reports, Ensemble uses an AI model to extract service details — the services required, minutes per week, frequency, setting, and provider type, plus the supplementary aids and supports listed in the same service tables. That's the extent of it. Every extraction is presented to district staff for review and confirmation before anything is saved; the AI drafts, people decide.
What AI never does. Student data is never used to train, tune, or improve any AI model — ours or anyone else's. The AI does not make decisions about students, does not score or rank students or staff, and does not generate assessments, evaluations, or recommendations about any child. Scheduling logic in Ensemble, including suggested schedule fixes, is deterministic software, not AI.
Documents aren't stored. Uploaded IEP files are processed in memory and never saved. Ensemble keeps only the minimum structured fields needed to schedule and log services.
Human review, always. AI-assisted output is a starting point for professional judgment, never a substitute for it. District staff review and validate extracted information before it enters a student's record.
If our use of AI ever expands, districts get written notice first. That commitment is contractual — it's in the Student Data Privacy and AI Addendum every pilot district signs.
The services behind Ensemble
Ensemble runs on a small number of infrastructure providers ("subprocessors"). Each one that touches student data is bound by written agreements consistent with our district commitments: data used only to provide the service, no advertising, no sale, and no use of student data to train or improve AI models.
Services that process student data:
- Anthropic — AI processing. Extracts service information from district-provided documents. Documents are processed in memory; student data is not used for model training.
- Supabase — database. Stores the structured scheduling and service-log records, isolated per district.
- Vercel — hosting. Runs the application.
Services that do not receive student data:
- Clerk — staff sign-in. Holds district staff accounts only; no student information is ever stored in the authentication system.
Our commitments, in one place
Districts own their data and can export or delete it at any time. Ensemble holds the minimum data needed to do its job, under per-district isolation. Student data is never sold, never shared for advertising, and never used for any purpose other than serving that district. The full terms live in the Pilot MOU and the Student Data Privacy and AI Addendum — ask us for a copy anytime.
SeaChi Group LLC, which operates Ensemble Scheduler, operates within FERPA’s school official framework, helping districts deliver educational services while maintaining control over student information. We access and use education records only for district-authorized purposes. Student data is not sold, used for advertising, or used to train any AI models. Our access to and use of student data is governed by our agreements with participating districts, including the Student Data Privacy and AI Addendum.
For how we handle information collected through our website, see our privacy statement.