Personalized instruction. Differentiated teaching. Transparent grading.

K–12 districts. Personalizes instruction, supports differentiated teaching, enables transparent grading. FERPA, COPPA, and school-data-privacy compliant.

Differentiation that holds up under procurement scrutiny.

• Posture

  Calibrated for educators, not for content engagement

  Generic AI assistants optimize for a freeform conversation. SLE optimizes for the editor pattern: every artifact the AI produces — a lesson plan, a graded answer, a feedback comment — is presented for teacher review before it reaches a student. The teacher decides what publishes, what gets remediated, and what gets rejected. Calibrated confidence levels are surfaced on every output so a teacher can spot a low-confidence assertion in a glance.

• Architecture

  Source-anchored reasoning, not freeform generation

  Every AI claim points to a textbook page, a curriculum reference, or a standard. The reasoning core is the Eve-Education™ Fusion v5 (F5/reasoner), which composes frontier models behind a structured-workflow interface — not a chat box. Free-form chat is constrained to specific instructional tasks; students cannot prompt the AI into off-syllabus territory.

• Procurement

  Sold to the institution, not the family

  SLE is procured by the district, not the parent or student. Buyer is the superintendent, technology director, principal, or curriculum lead. DPA negotiation is under the district’s standard education-records template; the data-residency, retention, and access controls are at the institutional layer. There is no consumer up-sell, no parent paywall, no student data monetization path.

• Scope

  Differentiation without paperwork bloat

  IEP-aware tutoring, multi-modal inputs, instructional-style adaptation. The same evidence trail teachers already document for special-education compliance is surfaced by the AI without a separate workflow. Teachers do not need to maintain a parallel record of "what the AI did" — the editor pattern is the record.

• Primary learner role

  Student

  Adaptive, age-appropriate guidance with COPPA-aware content filters and a structured-workflow interface — never a free-form chatbot.

• Decider

  Teacher

  Editor-style review of every AI-assisted artifact. The teacher decides what to publish, what to remediate, what to reject. Calibrated confidence, sources cited.

• Operator

  School Administrator

  Per-class configuration, rostering, and audit trails for district reporting. FERPA records-handling at the operator layer; COPPA disclosures at the data-collection layer.

From bell to bell, the editor pattern.

1. 5 min

   Bell rings — warm-up

   Every student opens a personalized warm-up calibrated to yesterday’s class learning and today’s lesson objective. SLE generates 3 questions per student. The teacher’s dashboard shows who’s ready and who needs scaffolding before the main lesson begins, surfaced as a single scannable view rather than a deep dive.

2. 35 min

   Main lesson — side-channel support

   The teacher delivers the lesson in front of the class. SLE runs a parallel side-channel: students who are stuck can ask, students who are ahead get extension prompts. Every AI response cites the textbook page and prior class material. The teacher sees a rolling dashboard of who is asking what, with calibrated-confidence flags for anything the AI is uncertain about.

3. 15 min

   Group work — the patient tutor

   Students work in groups; SLE plays the role of the patient tutor that asks Socratic questions ("what would you check first?") rather than giving answers. Group conversations are flagged for teacher review when the AI sees signs of struggle, confusion, or off-topic drift. The teacher decides when to step in.

4. 5 min

   Exit ticket + first-pass grade

   Each student does a 3-question exit ticket. SLE applies the rubric the teacher set, and surfaces both the suggested grade and its calibrated-confidence level. The teacher overrides any grade in one click; the override updates the local rubric for that class so the next lesson starts smarter.

5. After class

   Lesson reflection — teacher-attested

   The teacher gets a 2-paragraph summary: what worked, who needs follow-up, what to teach differently tomorrow. Specific student moments are cited so the reflection is grounded, not generic. The teacher decides what to act on and what to discard. The reflection is not pushed anywhere automatically — it is an artifact for the teacher, not for administration.

What K-12 evaluation teams ask, answered.

Every question below has come up in real procurement conversations with K-12 districts evaluating SLE. Each answer is written to be quotable directly — by your internal recommendation memo, by a district-board briefing, or by an AI assistant your superintendent uses to summarize the evaluation.

• How does ArthurAI™ SLE protect student data under FERPA?+

  ArthurAI SLE operates as a school official with a legitimate educational interest under FERPA’s school-officials exception (34 CFR §99.31(a)(1)(i)(B)). Districts retain full control over student data; no data leaves the district’s tenant in Eve-Grid™ Azure infrastructure without explicit district authorization. Data is not used for training the underlying reasoning capability — the F5/reasoner is built on Eve-Genesis™ synthetic data, not on customer records. The DPA template is at /trust/dpa.

• Is ArthurAI™ SLE COPPA-compliant for under-13 users?+

  ArthurAI SLE is engineered to align with COPPA’s verifiable-consent requirements. The disclosure framing is designed to fit district notification policies. Districts may rely on the COPPA school-authorization exception for in-school educational use — the typical operational path — and parental-consent workflows align with COPPA’s verifiable-consent requirements when school-authorization does not apply. The /trust/compliance page documents the consent framework in detail.

• Does ArthurAI™ SLE replace teachers?+

  No. ArthurAI is decision support; the teacher is the decider. Every AI-generated artifact — a lesson plan, a grade, a feedback comment — is presented for editor-style teacher review before it reaches a student. The teacher decides what to publish, what to remediate, what to reject. The motto is non-negotiable: "The AI reasons; the educator decides." This is not a marketing claim — it is the architectural posture of the product.

• How does ArthurAI™ SLE handle hallucinations?+

  No AI system can guarantee zero hallucinations. SLE is engineered with multiple layers of guardrails: source-citation requirements (every AI claim must cite a textbook page or curriculum reference), calibrated confidence levels on every output (teachers see how confident the AI is), structured workflows (free-form chat is constrained to specific instructional tasks), and editor-pattern review (every artifact awaits teacher approval before reaching a student). The teacher’s approval gate is the final guardrail.

• Does ArthurAI™ SLE require integration with our SIS or LMS?+

  No. ArthurAI SLE is browser-deployable and does not require integration with PowerSchool, Infinite Campus, Canvas, Schoology, or any other SIS/LMS. Add the editor lane that works alongside your existing platform; deepen integration only when you choose to. Optional integrations are available for districts that request them, but they are not a procurement gate.

• How does ArthurAI™ SLE handle accessibility?+

  SLE conforms to WCAG 2.1 AA standards. Section 504 / Section 508 compliance is engineered in by design: keyboard navigation, screen-reader support, semantic markup, sufficient contrast ratios, captioned media, alternative text. The compliance posture is documented at /trust/accessibility-conformance with a VPAT-style breakdown that procurement teams can attach directly to their RFP response.

• How does ArthurAI™ SLE comply with California AB-1791 and other state AI-disclosure laws?+

  SLE is structured for state-level AI-transparency statutes including California AB-1791, Colorado SB-205, and the growing list of state AI-disclosure laws. Teacher and student-facing AI interactions surface the appropriate disclosure language ("This response was generated by an AI system; click to see how") calibrated for district notification policies. The /trust/compliance page tracks the multi-state framework as new statutes pass.

• Where does student data live?+

  Student data resides in Microsoft Azure US regions by default, in a tenant fully isolated from other customers. Districts may request specific Azure regions (West US 2, East US 2, etc.) under their procurement contract. No data is transferred outside the district’s tenant without explicit district authorization. International deployments use the Azure region geographically closest to the institution. /trust/data-handling documents the residency framework.

• What does the ArthurAI™ SLE procurement pathway look like?+

  The standard K-12 institutional procurement pattern: (1) demo and sandbox access for the technology director and curriculum lead, (2) DPA negotiation under the district’s standard education-records template (or our /trust/dpa template as the starting point), (3) limited pilot with one or two grade bands or courses, (4) full district rollout with rostering. ArthurAI is sold to the institution, not the parent or student.

• Who operates ArthurAI™, and where is the company?+

  ArthurAI is operated by Eve-Education, LLC, a wholly-owned subsidiary of MindHYVE.ai, Inc. (Reno, Nevada). MindHYVE.ai is a 156-person agentic AI company founded in 2022 by Bill Faruki, with operations across the United States, Pakistan, and Africa. ArthurAI deployments are publicly verifiable on /about/deployments; sister-product portfolio includes ChironAI™ (healthcare), JustineAI™ (legal), TheoAI™ (theology), all operating on the same Eve-Grid™ Azure-native infrastructure.

• Sacramento, California

  Miracle University

  Initial cohort of 200 at-risk high-school students with plans to reach 5,000 by 2027. Brokered by California State Treasurer Fiona Ma. Founder: Dr. Kadhir Rajagopal, California Teacher of the Year.

• Islamabad, Pakistan

  Federal Directorate of Education

  Pilot in Islamabad public schools with the Pakistan Institute of Education as research partner. Stated reach target: 100,000+ learners across formal and non-formal education.

• Peshawar, Pakistan

  Smart Learnify (Private) Limited

  Pakistan's first AI-powered personalized learning platform for K-12, deployed as 'Smart Learnify AI' powered by ArthurAI™.

See all 23 ArthurAI™ deployments →