Problem Brief · Concept / Discovery

Proactive Teacher Alerts

Impacter Pathway Status: Concept / Discovery Author: Gary Hatton Sep 2026

1Problem Statement

Impacter Pathway scores durable-skill evidence from student voice responses and makes that evidence available to educators through a dashboard. However, the platform relies on a pull model — a teacher has to know to go look, and remember to check regularly, in order to notice a student who needs attention. Given typical teacher workload and the number of students per roster, this creates a real risk that meaningful signal (a declining Growth Mindset trend, a sudden drop in Purpose language, evidence of a student disengaging) goes unseen until it's already a larger problem.

This isn't a data or scoring problem — Impacter's core measurement is the differentiator. It's a workflow adoption problem: the best evidence in the world doesn't help a student if the person who could act on it never sees it in time.

2Why Now / Why It Matters

Impacter's core value proposition rests on educators actually acting on the evidence the platform surfaces. If evidence sits unseen in a dashboard, the product's ROI story to districts weakens regardless of scoring accuracy.
This is also a retention/renewal risk: districts renew tools that visibly change teacher behavior and student outcomes, not tools that produce reports nobody opens.
Competitively, most SEL/assessment tools in this market are already dashboard-first and pull-based (Panorama, Aperture, etc.) — a push-based intervention layer would be a meaningful differentiator, not a "me too" feature.

3Who This Affects

Primary: Classroom teachers Secondary: Admins / counselors Indirect: District renewal decision-makers

Teachers are expected to act on evidence but have the least time to go looking for it. Admins/counselors may want visibility across multiple classrooms. Districts evaluating renewal care whether the tool demonstrably changes outcomes, not just whether scoring is accurate.

4Current State (assumed, pending validation)

Teachers can view scores and evidence per student, per response, but must navigate to find it.
No existing mechanism proactively surfaces which students most need attention, or why, without a teacher going looking.
Unknown — needs discovery: how often teachers currently check the dashboard, and what would make an alert trustworthy enough to act on without over-alerting and causing fatigue.

5Proposed Direction

Surface students whose evidence indicates a meaningful change (decline, sustained dip, missing signal) without requiring the teacher to browse.
Tie every alert to specific evidence — the exact phrase and rubric level that triggered it — so it reads as decision support, not a black-box judgment.
Suggest a concrete next step, not just a flag, so the teacher isn't left with "now what."
Give teachers/admins control over alert sensitivity and which skills to watch, so trust and adoption aren't undermined by alert fatigue or a feeling of being surveilled by an algorithm.
Close the loop — let a teacher mark an alert addressed, add a note, or schedule a follow-up, so the system can also report back on whether interventions are working ("improved since last week").

6Risks / Open Questions

⚑

Alert fatigueIf thresholds are too sensitive, teachers will learn to ignore alerts entirely, defeating the purpose. Needs real tuning with actual teacher feedback, not just engineering defaults.

⚖︎

Fairness / biasLanguage-based scoring may behave differently across dialects, ELL students, or cultural communication styles. An alert system amplifies whatever bias exists in the underlying scoring — this needs explicit validation before launch, not after.

🤝

TrustA wrong or overly clinical-sounding alert about a student's mental/emotional state could damage teacher trust in the whole platform. Suggested next steps need to read as decision support, never as a diagnosis.

🧭

Scope of "intervention"Need to define with Impacter's team how prescriptive suggested actions should be — is this coaching language, or does it risk overstepping into territory better handled by school counselors/psychologists?

7Success Metrics (hypothesis — to validate)

📬 Primary KPI 1 · Response rate

% of alerts opened within 48 hours of generation

Target: 70%+

✅ Primary KPI 2 · Resolution rate

% of alerts marked "addressed" vs. ignored/expired

Target: 60%+

📈 Primary KPI 3 · Intervention effectiveness

% of flagged students who showed measurable skill improvement within 2 responses of an intervention

Target: 50%+

Supporting metrics:

Teacher-reported trust/usefulness score (qualitative + survey)

Improvement vs. baseline — flagged-student improvement rate compared to a comparable unflagged control group, to confirm gains aren't just natural fluctuation

8Not Doing (explicitly out of scope for v1)

No facial/biometric signal of any kind (privacy, trust, and regulatory risk far outweigh unclear signal gain).
No fully automated interventions — a human (teacher) must remain in the loop for every action taken.
No parent-facing version of alerts in v1 — start with the teacher workflow, expand later once the core loop is validated.

9Next Steps

Validate the core assumption with 3-5 real teachers: do they currently miss signal that a pull-based dashboard should have surfaced?

Prototype the alert feed + detail view (done — see teacher-alert-prototype.html) to pressure-test the concept in a stakeholder conversation before committing engineering time.

Define alert-threshold defaults with Impacter's data/psychometrics team, not just product intuition — this is a fairness-sensitive decision.