
2 in 5 high schoolers are quietly struggling, and no score in any school system flags them. The warning signs sit unread in the school's own records. Pebble is the number that finds those kids before the crisis.
high schoolers report persistent sadness or hopelessness. Suicide is the 2nd leading cause of death, ages 10-24.CDC YRBS 2023 · NCHS 2024
to assemble one student's picture across disconnected systems — attendance here, grades there, behavior somewhere else. So it happens after the crisis.
the warning signs are logged into the SIS — and nobody reads them. The status quo is a spreadsheet, a gut feeling, and whoever knocks on the door.
Real product · student profileAttendance, grades, and behavior — the triangulation a great counselor does by hand, run continuously across the whole caseload.
Drafts the profile, intervention options, and the MTSS meeting agenda. The counselor decides; the software prepares.
No new surveys, no data entry. Plugs into the student-records system the district already runs (Infinite Campus live; Mindex partnership signed). Kids are surveyed out; passive data is honest.
The people in the building20+ years in education. Founded and ran an alternative high school for at-risk students inside his district. Trained counselor.
20+ years alongside Scott, classroom to instructional technology. Owns SIS integration and district procurement.
Built and scaled companies in the education space. Runs finance and the raise.
Ex-Google Cloud staff architect, 30 years building enterprise-scale platforms — the last decade in cloud and AI for highly regulated industries. Led Pebble's build to date.
3,000 students, 4 buildings. Director of Pupil Personnel Services championing inside.
A BOCES network (New York's shared-service agencies that buy on behalf of many districts at once) plus the Westchester County mental-health department (commissioner engaged, May 2026).
Great Neck NY · Hanover SD · NYC DOE (two procurement paths) · founders' home district internal pilot.
Privacy Policy, ToS, district DPAs — already in front of district counsel. In K-12, procurement-ready is traction.
Real product · caseload
Congress appropriated $1B+ for school mental-health staff, and states are distributing $50B+ in opioid-settlement funds that can fund youth prevention. Districts can buy Pebble without touching classroom budgets.
~74% of schools now run "MTSS": the standard routine where staff meet over data to spot and support struggling students. The process is universal; the software for it doesn't exist. That's the hole Pebble fills.
~1 in 4 students is chronically absent, still far above pre-pandemic levels. Districts are publicly measured on it, and attendance is Pebble's strongest early-warning signal.
The systems that hold grades and attendance now expose modern integration standards, making a product like Pebble buildable by a small team for the first time.
One mid-size district (~3,000 students). NY's ratio is 331:1 — the beachhead state has the problem and the budgets.
BOCES + county buyers mean channel sales, not door-to-door. $5K paid pilots open doors where procurement wants a small first step.
Analyst estimates of the SEL/student-wellness software market, most projecting 20%+ growth. Our math is bottoms-up; the analysts just agree it's big.

Per enrolled student, billed annually — a 3,000-student district.
Single school, where procurement wants a small first step (the NYC DOE path).
Many districts in one motion. The median public-school counselor costs $77K (BLS) — Pebble prices far below one hire and multiplies the ones they have.
Pre-revenue today, and we say so. What we have is pricing districts told us "nobody blinks at."


Districts ask about privacy before they ask about price, and an unprepared vendor loses months in legal review. We did this work up front, so pilots start instead of stalling.
NY districts, through the Brewster pilot and the BOCES / Westchester channels.
Every MTSS meeting (the data-review huddles 3 in 4 schools already run) planned, tracked, and measured through Pebble.
Anonymized, consented data across districts: education's first longitudinal picture of student wellbeing.
Brewster live — 3,000 students, 4 buildings. The lighthouse reference.
First paid contracts · channel deals opening through BOCES & Westchester.
10-20 districts · $1-2M ARR target — 30-60K students at $30-35, the same math as our pricing · SOC 2 Type II · Series A raised on pilot data, not projections.
A1 How the concern index works · A2 Privacy & security · A3 Pipeline detail · A4 AI governance
| Element | What it does | Why it's built this way |
|---|---|---|
| Inputs | Attendance patterns, grade movement, and behavior incidents, pulled nightly from the district's student-records system. | Passive data only. No surveys, no new data entry, nothing for kids to game or skip. |
| Scoring | Each student gets a current-quarter concern score; high = needs attention. Sub-scores per signal stay visible so counselors see why. | Counselors told us the "current quarter" snapshot is the working view; yearly averages hide the kid who fell off a cliff in March. |
| Trends | Longitudinal view per student (quarter, year, career) behind one toggle. | The pattern over time is the conversation starter with parents and MTSS teams. |
| Validation | Weighting tuned and checked against real student records in pilot work, with practitioner review of every iteration. | We publish no accuracy percentages until we have peer-reviewable pilot data. Targets stay labeled as targets. |
| Human in the loop | The index ranks and explains. It never diagnoses, never labels a student, never triggers an automated action. | Counselors decide. Always. |
| Area | Status |
|---|---|
| NY Ed Law 2-d | Architected for the strictest US state student-privacy law: US data residency, annual third-party security audit path, parent data rights. NY-compliant means 50-state-ready. |
| FERPA | "School official" standard enforced in code: every user's access scoped to their caseload; every student-record access logged. |
| Security framework | 269 NIST CSF controls mapped and continuously monitored in Drata. SOC 2 Type II completion funded by this round. |
| Legal stack | Privacy Policy, Terms of Service, and district data-privacy agreements drafted by Kirkland & Ellis; in front of district counsel now. |
| Data boundaries | Student data is never sold, never used to train third-party models, never leaves US infrastructure. Vendor data-processing agreements in place or in progress for every subprocessor. |
| System | Scale | Stage · honest label |
|---|---|---|
| Brewster CSD (NY) | ~3,000 students · 4 buildings | Pilot launching Sept 2026. Director of Pupil Personnel Services championing; legal docs with district counsel. |
| Founders' home district | District-wide | Internal pilot environment ready; staff relationships in place. |
| BOCES network (NY) | 25+ districts via one agency | Warm introduction made; channel conversation. BOCES = NY's shared-service agencies that purchase for member districts. |
| Westchester County Dept. of Community Mental Health | County-level buyer | Commissioner engaged May 2026; follow-up call scheduled. County funding streams (federal grants, opioid settlement) can fund district adoption. |
| NYC DOE | Largest US district | Two procurement paths identified: $5K/school paid pilots now, or approved-vendor track at $25K+/school. Early conversations. |
| Great Neck (NY) · Hanover SD | District | Demos held; conversations active, no commitments yet. |
| Principle | In practice |
|---|---|
| Co-pilot, not decision-maker | AI drafts student summaries, intervention options, and meeting agendas. A counselor reviews everything before it touches a student's life. No automated decisions about children, ever. |
| Scoring is transparent math | The concern index is deterministic weighting over school records, not a black-box model. Counselors can see exactly which signal moved a score. |
| Data minimization | AI features receive only the fields needed for the task, under data-processing agreements; student data is never used to train foundation models. |
| Auditability | Every AI-generated draft is logged with its inputs; districts can review any output's provenance. |
| Positioning with skeptics | We pitch to educators who distrust AI, deliberately. The product wins them by preparing their work, never replacing their judgment. |