For facilities & platform partners

Control is part of the product.

A useful correctional learning application has to answer more than “What can the AI do?” It must make content, identity, data, deployment, oversight, updates, refusal, and shutdown decisions inspectable.

Abbé Faria invites technical-fit conversations and focused evaluations designed around institutional requirements.

The control stack

Every capability needs an owner.

Authority is visible from the first conversation, giving facilities and partners a clear foundation for evaluation.

  1. 01ContentApproved sources and lessons
  2. 02IdentityAccess and learner roles
  3. 03RuntimeDevice and network limits
  4. 04OversightReview and escalation
  5. 05ContinuityRecovery and rollback

Governance model

Each party keeps the authority it should have.

FacilityApproves availability, content, oversight, learner support, data retention, and administrative controls.
Platform partnerControls distribution, device policy, identity, facility access, approved runtime, and integration pathway.
Abbé FariaProvides the learning experience, curriculum workspace, bounded model behavior, source-linked answers, evaluation, and release package.
Shared decisionDefines the pilot cohort, success measures, support plan, accessibility review, decision points, escalation, rollback, and data disposition.

Proposed evaluation path

Fit first. Evidence next. Expansion only when earned.

  1. 01Discover

    Confirm the target device, runtime, network boundary, identity, review, and procurement pathway.

  2. 02Define

    Select one approved learning objective, participant criteria, data rules, success measures, and completion criteria.

  3. 03Verify

    Validate source support, scoped responses, privacy, accessibility, updates, recovery, network behavior, and administrative control.

  4. 04Pilot

    Run a focused, voluntary evaluation with platform, facility, content, security, support, and measurement alignment.

Decision-maker checklist

Questions worth answering before a pilot.

Educational fit

  • Who are the intended learners?
  • Which skills and approved sources?
  • What counts as mastery?
  • How will educators review and revise content?

Safety & well-being

  • Which requests need specialist support or escalation?
  • How are care boundaries communicated?
  • What accessibility and language review is required?
  • How is learner privacy protected?

Operational control

  • Where should inference run?
  • How should identity, retention, and updates be governed?
  • Which operational reports support review?
  • What controls are required for rollback and closure?

A technical-fit conversation is the next useful step.