Meet Sammy.

Everyone from ages 2 to 92 — learners, families, teachers, specialists, and school leaders. Sammy adapts to whoever she's paired with: a fifth grader building reading confidence, a parent organizing family life, a teacher planning small-group instruction, or an adult chasing personal goals. And because she pairs for the full lifecycle — ages 2 to 92 — the bond simply changes shape as life does: reading confidence at eight, study systems at eighteen, career momentum at forty, memory and routine support at eighty.

Agentic Companions are a new category of intelligence, created by Agentic Learning and Sammy — specialized supports she brings into the moment — like the ADHD Agentic Companion, Dyslexia Agentic Companion, Behavior Agentic Companion, or Emerging Bilingual Texas Sammy. They all run on one Master Kernel, so they share one memory and one truth engine. Parents and schools choose which Agentic Companions are enabled, and access can be changed at any time. Every Agentic Companion runs on the same Master Kernel and permissioned memory, so turning one on adds expertise to the Sammy you already know — never another app to learn.

Yes. Sammy runs on permissioned memory: she only stores and uses what you allow. Topics can be set to On, Limited, or Off; each learner receives appropriate access; and your data is never sold — ever. It is held in industry-leading, SOC 2 Type II–level AI architecture engineered so HIPAA and FERPA compliance are managed at 100%, at all times. You're in control — always. You can export or delete your data on request, access logs show exactly which adults saw what, and nothing in the family zone is visible to a school without explicit, revocable consent.

Sammy connects home and school as one coordinated experience. Educators get workspaces for instruction planning, progress monitoring, and team collaboration; families see updates, routines, and wins; and shared visibility is governed by appropriate permissions. Schools and districts can book a demo to launch campus-wide. Home and school share one longitudinal record under permissioned control, which means the evidence a family gathers at the kitchen table arrives meeting-ready in the ARD room.

Sammy is built on an AI-native workflow architecture — a Master Kernel that combines a truth engine, permissioned memory, and an orchestration layer. She doesn't just answer; she verifies what is known, coordinates specialized Agentic Companions, and recommends the next best action, with explainable reasoning behind every suggestion. The deeper difference is governance: a chatbot answers and forgets, while Sammy is authorized on entry, verified on exit, and recorded permanently — accountable for every recommendation she makes.

Families and individuals start with a free 7-day trial, then it's just $5/month — cancel anytime. Schools and districts get tailored plans with workspaces, dashboards, and admin controls; book a demo and we'll walk you through it. The full ladder: $5/month for families after the free 7-day trial, $49 at the individual professional tier with every switch free to flip, $49–$89 per student per year for districts by volume, and one-time products — IPCCA, ENOUGH!, Bridge — bought once and owned forever.

Every Sammy Agentic Companion wears her colors. Nine hanky colors map to areas of need, support, and compliance — purple for the Autism / Continuum Agentic Companion, orange for ADHD & Executive Function, red-orange for Dyslexia / Dysgraphia / Dyscalculia, blue for Communication & Language, teal for Behavior & Emotional Support, gold for Life Skills / IDD / Transition, cyan for Emergent Bilingual, rainbow for Gender, Identity & Belonging, and deep navy for Compliance Areas. One glance tells you which support Sammy is showing up for. Hankies layer like the supports themselves do — and the deep-navy compliance hanky is always underneath, because governance never comes off.

Individuals and families start with a free 7-day trial, then just $5/month. Beyond that, Sammy scales to every setting: therapy team licenses, school site licenses, district enterprise agreements, professional development for educators and teams, expert services and implementation, and API / white-label options to embed Sammy in your own platform. Progressive pricing runs through all of it: every dashboard or add-on you enable costs less than the last, packs beat à-la-carte when usage grows, and personal purchases roll up automatically when a district signs.

Not at all — Sammy is an Agentic Companion for life, from ages 2 to 92. Beyond classrooms, she organizes family life, coaches personal goals, supports college and lifelong learners, and even helps pet parents with insight, planning, memory, and guidance for everyday decisions. Think clearly. Stay organized. Act confidently. One Agentic Companion. One ecosystem. Endless possibilities. The same paw prints follow you from a third-grade reading goal to a college application, a first job, a growing family, and a well-organized retirement — one record, one bond, the full lifecycle.

District pricing is volume-based. Agentic IEP — the flagship IEP management and writing platform with the full resource kit and a GPT & agent toolkit customized to every user — runs $49 per student for the largest districts and $89 per student for the smallest, with fair tiers for everything in between. Once any dashboard or bundle is purchased, every additional feature and function unlocks at progressively lower cost. Districts can turn on every access point from day one and simply true up at the end of the contract year or cycle — and parents can do the same. Best of all, nobody waits: teachers, students, and admins can start today with a credit card via a simple QR code, and their accounts roll up automatically when the district signs. And districts themselves can move just as fast: subscriptions can be purchased directly with a district credit card — in seconds — with the district up and running immediately, while procurement formalizes the contract and trues up later. The tier builder arrives pre-checked with logged teacher requests, prices the selection live against the volume bands ($77–$89 under 5,000 students down to $49–$55 above 100,000), and produces the quote, PO paperwork, and security review packet in a single pass.

The Agentic Learning Algo is the proprietary measurement and cognitive engine that powers every Agentic product, Sammy included. It models learning as a rising spiral rather than a straight line, holds every recommendation to measurement-grade evidence — valid baselines, aimlines, and psychometrically sound progress interpretation — and orchestrates a constellation of 40 specialized agents covering every legal, educational, and assessment requirement. The component frameworks beneath it are proprietary; publicly, the Agentic Learning Algo is all you need to know. In practice that means learning modeled as a rising spiral rather than a line, a constellation of forty specialized agents covering every legal, educational, and assessment requirement, and measurement-grade discipline behind every number you see.

The house assessment exemplar, the Agentic-AS, was field-tested with 740 participants — 420 youth in grades 4–12 and 320 adults ages 18–65 — stratified by age, gender, race/ethnicity, region, SES, language background, and neurodiversity status. Construct validity: confirmatory factor analysis supports the five-domain structure (CFI = .96, RMSEA = .036). Content validity: an eight-member expert panel produced an average item-objective congruence of .91. Concurrent validity: r = .73 with WISC-V Full Scale IQ and r = .72 with performance-based reasoning on Woodcock-Johnson IV and WAIS-IV scales. Predictive validity: cumulative GPA r = .68, and success in real-world innovation challenges at an odds ratio of 2.8 (p < .01). And every coefficient carries full dossier provenance — the criterion measure, sample size, confidence interval, and evidence date — so the record is reviewable, not folkloric.

Every coefficient clears industry thresholds with room to spare: internal consistency α = .92 (domain range .87–.94, against a ≥.80 standard), six-month test-retest ICC = .87, inter-rater ICC = .90, and a generalizability coefficient of .89 against a ≥.85 standard. Reliability isn't an afterthought here — it's the entry ticket. The generalizability coefficient of .89 matters most in practice: it means scores hold steady across raters and occasions, which is exactly what a decision-grade measure must do.

By taxonomy, not convenience. Tier 1 — norm-referenced psychometric anchors — authorizes. Tier 2 — standardized, progress-monitorable measures — operates day to day. Tier 3 — grades and classroom judgment — corroborates only, and never serves as a baseline. Five disqualification gates prevent a bare standard score from ever anchoring a goal it can't support, and platform estimates follow a provisional → burn-in → lock lifecycle before they're trusted. The full step-by-step walk-through lives in the Baselines & the Tiered System category below — from candidate data to locked anchor to aimline.

A rule with teeth: every cross-instrument correlation used in concordance-band math must carry full provenance — the coefficient itself, the criterion measure, the sample size, the confidence interval, and the date of the evidence. No orphan correlations, no folklore coefficients. If a number can't show its dossier, it doesn't get to move a decision. A dossier row reads like a lab record: r = .72, criterion = performance-based reasoning composite, n = 740, 95% CI reported, evidence dated — every element present or the coefficient sits out.

Through the Ambition Quotient — an operationalization of the Endrew F. "appropriately ambitious" standard. Goals are set against measurement-grade baselines and aimlines, never vibes, and progress is interpreted the way a psychometrician would read it. Longitudinal data is preserved, not flattened: a conflict exists only when the same assessment reports different scores on the same date — growth across dates is a story, not an error. And the aimline is re-read every monitoring cycle: when growth beats the line, the system recalibrates upward rather than coasting — ambition is maintained, not just declared.

Yes — that's the design. Norm-referenced instruments serve as Tier 1 anchors: the platform consumes scores and interprets them, and never reproduces items, stimuli, or protected content. It is fluent in every major score scale — standard scores (M = 100, SD = 15), scaled scores (M = 10, SD = 3), T-scores (M = 50, SD = 10), and NCEs for program-evaluation reporting — so your instruments arrive speaking their own language and leave with their security intact. No item content, stimuli, or protected material is ever stored or reproduced — the platform is a score consumer and interpreter, which keeps publisher IP exactly where it belongs.

Built for it. Technical adequacy documentation, concurrent validity dossiers, and the permanent Evidence Ledger are structured so a third-party review team can trace any output back to its inputs, rules, and sources. Anti-fabrication protocols arm before reasoning begins, and every claim in every document cites the evidence beneath it. Reviewers can request an Evidence Ledger export for any pass: inputs, rules fired, gates cleared, and outputs — the complete chain of custody for a decision.

A psychometrician, first. The founder holds a doctorate in measurement & evaluation from the University of Virginia and conducted psychometrics research at Pearson Clinical — and the platform inherits that posture: a psychometrician-in-the-loop philosophy governs how scores are anchored, interpreted, and allowed to move decisions. Interpretation rules are encoded, not improvised: confidence intervals respected, base rates considered, and no single score ever asked to carry more weight than its reliability can bear.

Yes — API and white-label licensing is a first-class delivery line. Partner consoles inherit the same gated AIMS architecture: Human Authority Gates on entry, Verifier Agents on exit, the Evidence Ledger underneath, and permissioned memory throughout. You bring the surface; the engine brings the discipline. Integrations stage cleanly — read-only evidence services first, gated generation second, full workflow orchestration third — so partners ship value early and deepen on proof.

An AI-native, workflow-driven architecture: twelve kernels, five mandatory governance structures, and a forty-four-waypoint thinking path where every rule is visible and every gate is enforced before reasoning, not audited after. The Agentic Learning Algo powers cognition as a proprietary engine with explainable outputs — partners integrate its results and its reasoning trail, not its internals. At full resolution the pass runs forty-four visible waypoints through the five mandatory structures — Human Authority Gates, Verifier Agents, Continuous Oversight, the Evidence Ledger, and Longitudinal State.

Cleanly. The component frameworks beneath the Agentic Learning Algo remain proprietary; partners license capabilities, interfaces, and outcomes. Role-based access, audit trails, and contractual guardrails sit at every seam — so partnership accelerates both sides without blurring what belongs to whom. The engine is licensed like an engine should be: partners get its power, its outputs, and its explainability — the proprietary frameworks beneath it stay home.

It was architected for exactly that. The same pass loads a different jurisdictional ruleset — IDEA floor, agency rules, local policy — without the architecture changing, and SOC 2 Type II–level controls, encryption, role-based access, and audit trails hold from one educator to a mega district to an education agency. Scale changes the quote, never the rigor. The proof is live: a district-wide production deployment across 276+ campuses and roughly 30,000 students runs through the identical gated pass a single classroom uses.

Through Agentic EB and the TELPAS-aligned Emerging Bilingual Texas Sammy: language growth tracked across listening, speaking, reading, and writing; LPAC and ARD–LPAC workflows streamlined; family communication translated in real time; and reclassification readiness projected program-wide. Multilingual learners get celebrated, and programs get the evidence their outcomes deserve. Growth is tracked to jurisdictional proficiency descriptors across all four language domains, and reclassification readiness is projected learner by learner, campus by campus.

Yes. The platform is score-scale fluent across instruments, performs cross-instrument concordance only with dossier-backed concurrent validity coefficients, and reports in the conventions program evaluators expect — including NCEs for program reporting. Your ecosystem stays the system of record; the platform makes its data work harder. Concordance across instruments happens only through dossier-backed concurrent validity coefficients — never assumed equivalence — so cross-ecosystem numbers stay defensible.

The Agentic-AS measures five domains confirmed by factor analysis: analytical reasoning, conceptual synthesis, strategic questioning (treating powerful inquiry as a core skill, not a disposition), AI-augmented problem solving (leveraging tools as collaborative agents in original intellectual work), and iterative adaptation (revising and improving through recursive learning cycles). Confirmatory factor analysis supports the five-domain structure with χ² = 248.6, CFI = .96, and RMSEA = .036. Both forms measure the same five-domain construct — the Youth Version for grades 4–12 and the Advanced Version for adults — so growth reads continuously across the lifecycle.

Field testing spanned 740 participants across two forms — a Youth Version normed on 420 learners in grades 4–12, and an Advanced Version normed on 320 adults ages 18–65 — with stratification by age, gender, race/ethnicity, geographic region, socioeconomic status, language background, and neurodiversity status. The sampling frame was built to reflect the learners the platform actually serves, including the neurodiverse populations most instruments underrepresent. Stratification wasn’t decorative: it spanned age, gender, race/ethnicity, geographic region, socioeconomic status, language background, and neurodiversity status — seven variables, held simultaneously.

Proven at district scale: a live, district-wide deployment spans 276+ campuses serving roughly 30,000 students, running at an 88.1% cumulative pipeline success rate with an average end-to-end runtime of about three and a half minutes per full document pass. Not a demo, not a pilot deck — production traffic through the full gated architecture, every day. Every one of those passes clears all five gates — authorized on entry, verified on exit, recorded permanently — at production speed.

Staged and clean: partners keep their surface, brand, and user relationship; the engine arrives underneath with the five mandatory structures intact. Integration proceeds from read-only evidence services, to gated generation, to full workflow orchestration — with explainable outputs and the complete reasoning trail exposed at every stage. The Agentic Learning Algo remains a licensed engine; the partnership ships outcomes, not internals. Partners keep their brand, their surface, and their user relationship end to end; the engine is invisible to their users and indispensable to their product.

Yes — Agentic EB is live as the platform's English-language vertical, with the TELPAS-aligned Texas edition leading the jurisdictional line: listening, speaking, reading, and writing growth tracked to jurisdictional proficiency descriptors; LPAC and ARD–LPAC coordination built in; real-time family language access; and reclassification analytics program-wide. The same architecture loads each jurisdiction's ruleset without rebuilding — Texas first, every jurisdiction next. Family communication arrives in the home language in real time, and LPAC documentation lands meeting-ready — the paperwork burden drops as the program evidence rises.

That's the operating condition it was built for. Procedural clocks arm before analysis begins, anti-fabrication protocols precede reasoning, every output passes Verifier Agents before release, and the Evidence Ledger records each decision with its rationale — under SOC 2 Type II–level controls, encryption, and role-based access. Peak season changes the volume; the gates never move. Procedural clocks arm the moment an event exists — before any analysis begins — which is precisely the discipline high-stakes seasons demand.

Your first dashboard is full rate; every one you add after it costs less than the last — for families and professionals alike. When your usage pattern means a pack would save you money, Sammy tells you before you'd notice yourself; you're never left paying à-la-carte prices for pack-level use. And progressive pricing survives organizational change: everything you unlocked personally carries forward as credit when your district signs.

Automatic roll-up. Your account, your settings, your history, and your flipped switches all move under the district license the day it activates; remaining personal-subscription value applies as credit; and there is no double-billing, ever. Early adopters are rewarded, never penalized — the teachers who bought their own toolkit are usually the exact people whose logged requests built the district's tier in the first place.

All three, matched to who you are. Families and individual educators pay by card or native billing — one gesture, expense-ready receipt, cancel anytime. Districts run the tier builder: components selected, priced live, quote and PO paperwork generated in one pass with the security review packet attached. One-time products are bought once and owned forever, and they follow the family across schools, districts, and years.

Yes — families get a full 7-day free trial with no credit card required to start, then $5/month, cancel anytime. The trial isn't a teaser: the Master Kernel, permissioned memory, and every consumer Agentic Companion your settings enable are live from the first minute. Educators can start the same day at the $49 individual tier, where every professional switch is free to flip.

The opposite — pricing rewards scale at every level. Districts slide down the volume bands from $89 toward $49 per student as enrollment grows; families see each added toggle cost less than the one before; and the turn-everything-on model means you deploy first and true up at cycle end, so access never waits on accounting. Growth makes Sammy cheaper per learner, by design.

By value delivered, not tokens burned. Usage is counted in completed workflow passes — a plan built, a document drafted, a determination supported — and it's visible to you in plain dashboards, never buried in a metering mystery. Usage signals work in one direction only: they surface savings, like suggesting an All-Access pack the moment à-la-carte spending crosses it.

No. The family subscription covers your Agentic Companion's daily life — conversations, check-ins, plans, memory, and growth — without per-use charges or surprise overages. A learner who talks to Sammy fifty times a day pays exactly what a learner who visits twice a week pays. Intelligence that punishes engagement would be intelligence pointed the wrong way.

Under governance, always: every pass is authorized on entry, disciplined in execution, validated on exit, and recorded permanently, with human-authority questions framed for humans and never answered by machines. Product improvement draws only on aggregate, de-identified patterns under the Data Covenant — your family's specifics improve your family's experience, and nothing else.

Implementation intelligence, not surveillance. Agentic Command shows adoption, workflow throughput, outcome trends, and compliance risk across campuses — the signals leaders need to support their people. What it never shows is the private family zone: home conversations, family journals, and personal check-ins are architecturally invisible to institutional dashboards, consent-gated at the root.

Never sold. Ever. No advertisers, no data brokers, no behavioral-ad profiles, no monetization of learner information in any form — revenue is subscriptions and licenses, full stop. The business model is the privacy policy: when families and districts are the customers, learners never have to be the product.

Industry-leading, SOC 2 Type II–level AI architecture: encryption at rest and in transit, role-based access enforced on every request, permanent audit trails, and hard architectural separation between the family zone and institutional workspaces. Any tool that would touch a child's data must clear a consent gate before it can even activate — protection is enforced by the same gates that govern Sammy's thinking.

You do. Families own their family-zone data outright; districts hold their educational records under FERPA; and Sammy acts as steward of both — never owner of either. Export and deletion are available on request, access logs are yours to inspect, and consent is revocable at any time. Stewardship means the data works for you and answers to you.

Architecturally, not aspirationally. Compliance rules load as executable constraints — the same way IDEA and jurisdictional rules load into the Thinking Path — so HIPAA and FERPA requirements are enforced during every operation rather than audited after it. Monitoring is continuous, the Evidence Ledger records the proof permanently, and 100% is the operating condition, not the annual goal.

Six moves, every time. One: candidate data arrives and is classified by tier — Tier 1 norm-referenced anchors, Tier 2 standardized progress-monitorable measures, Tier 3 classroom judgment. Two: only Tier 1 can authorize and only Tier 2 can operate; Tier 3 corroborates. Three: the five disqualification gates screen any score asking to anchor a goal. Four: platform estimates enter provisional status and earn trust through burn-in across monitoring cycles before they lock. Five: the aimline is drawn from locked baseline to ambitious goal. Six: the whole chain — sources, tiers, gates, dossiers — is recorded permanently. No step is skippable, including for Sammy herself.

Because grades are honest about many things and standardized about none of them. A grade blends achievement with effort, behavior, curriculum, and teacher judgment; it carries no norms, no reliability coefficients, and no defined scale — so it can corroborate a story the measurement tells, but it can never start one. Tier 3 evidence is respected as context and refused as anchor, which protects both the learner and the teacher who assigned the grade.

In plain terms, a score asking to anchor a goal must show that it is recent enough to describe the learner today, that it measures the construct the goal actually targets, that its reliability is adequate for a decision of this weight, that its provenance dossier is complete, and that the resulting goal can genuinely be progress-monitored. Fail any one and the score keeps its place in the record — but loses its claim on the anchor.

The lifecycle every platform-generated estimate must survive before it's trusted. It enters provisional — usable for conversation, barred from decisions. Through burn-in, it must demonstrate stability across successive monitoring cycles, agreeing with the external evidence around it. Only then does it lock as a decision-grade anchor — and if later data shows drift, it reopens. Sammy holds her own estimates to the same standard she holds everyone else's.

The aimline is the contract between baseline and goal: a visible trajectory the data must climb. Ambition is set by the Ambition Quotient against the Endrew F. standard — appropriately ambitious in light of the child's circumstances, quantified rather than felt. Honesty is enforced by the trend read: actual growth versus the line, every cycle, with longitudinal points preserved rather than flattened. Beat the line and the goal recalibrates upward; fall below and the team knows early, while there's still time to act.

She waits thirty days, exactly as promised at the gate. In that window you can export everything, or come back like nothing happened. After thirty days she's actually deleted — not archived, not anonymized, deleted. And she never says a sad word about it on the way out.

No. Not support, not engineering, not the founder. That isn't a policy we're being brave about — it's the account architecture. You are the customer, so you are the only reader. The day that stops being true, the five dollars was a lie.

A certification, never the companion. Scoped access to the records it already holds under law, plus a full audit trail of every artifact she produced about that work — and not one word of anything else. The certification detaches clean the day you leave, and your Sammy stays yours.

No, and she'll say so out loud. She names her limits, and if you're struggling in a way that needs a real person, she says so and helps you find one. She pushes you toward the humans in your life — she never handles that in-house.

Then she isn't Sammy. The name is yours to give, in minute four, and the naming is the first act of the relationship — not a settings field. Most people keep Sammy. Some don't. She answers to what you chose, either way.

She is yours to take with you or to end — and that includes the end. You can leave standing instructions: delete everything, or release chosen exports to chosen people. What never happens is a third option someone else picked for you.

Completely. HomeSchool OS turns Sammy into the family's school office: student records, attendance, and portfolio kept automatically as you teach; curriculum mapped to jurisdictional requirements; the daily rhythm, co-op days, and field trips on one calendar; tutors and therapists scheduled and invoiced; and transcripts, progress reports, and filing reminders ready whenever the jurisdiction asks. You remain the teacher and the school — she carries the clipboard, the filing cabinet, and the checkbook.

Everything a five-person office would: a complete student information system for the cohort, mixed-age scheduling and studio rotations, tuition drafted and reconciled with provider payouts queued, guide contracts and background-check tracking, a parent hub with updates and conference booking translated automatically, and a compliance binder that stays continuously current instead of being assembled in a panic. MicroSchool Command runs $79/month for up to 25 learners, then $5 per learner.

Yes — the Private School Suite is a full replacement or a graceful overlay: enrollment through transcript in one living record, master scheduling, attendance, gradebook and report cards, tuition with financial-aid awards and autopay, related-services coordination across contracted providers, whole-community communication translated and delivery-tracked, and jurisdictional reporting drafted from live records. It prices on the same volume band as districts — $49–$89 per student per year — because a school office deserves district-grade infrastructure at any size.

No — and that line is printed in her limits on purpose. Sammy manages, coordinates, documents, and bills; accreditation, umbrella programs, diplomas, and jurisdictional standing remain with you, your teachers, and the responsible bodies. She is the operating system, never the institution. And every learner inside the schoolhouse still has their own paired Sammy — read by no one else, and still theirs long after graduation.