Analysis · AI

The cycle read-out, written for humans.

When a survey cycle closes, an integrated Anthropic Claude panel reads every response and writes the summary your district leadership needs: executive summary, strengths, areas of concern, and prioritized recommendations, in plain language.

How it works

Cycle closes. Analysis ready.

When each survey cycle completes, the analysis job gathers final scores, gaps, and open-response themes across the district. The model receives suppressed, anonymized aggregates only—never raw responses, never student names. It writes the summary; you bring the judgment.

  • Executive summary, strengths, areas of concern, and recommendations
  • Generated per campus and districtwide at the close of every cycle
✦ Climate summary · District
Executive summary

Districtwide favorability rose 3 points this cycle, led by gains in Belonging & Wellbeing (+5) and Family Voice (+6). Six of eight domains now sit at or above 74% favorable.

Strengths
Academic Excellence remains the district's strongest domain at 84%, consistent across all three grade bands.
Family–Teacher Communication improved at 14 of 16 campuses; open responses credit the new conference scheduling.
Areas of concern
A 17-point family–student gap on Belonging at Central High has persisted for two consecutive cycles.
Staff Workload & Role Clarity (61) is the lowest score in the ecosystem and the strongest driver of Retention Intent.
Prioritized recommendations
1 · Pair Central High's belonging gap with its open-response themes before the August leadership retreat.
2 · Review campus-level Workload & Role Clarity scores with principals before staffing plans close.
Powered by Anthropic Claude
Privacy by design

Sees aggregates. Never students.

The model receives only suppressed, anonymized score aggregates. Survey respondent PII is never stored alongside response data for full anonymity and privacy assurance.

  • No raw text responses—only domain scores and flagged gaps
  • FERPA-first by architecture, not by policy alone

PII never stored with responses

Survey tokens are one-way hashed. There is no database column joining a student's name to their answers.

Per cycle, not always-on

The AI job runs once per cycle when it closes. There's no live model connection to the response database.

Wake up to the analysis, not the spreadsheet.

Start with a single campus cycle. We handle the instruments, the languages, the privacy floor, and the cycle summary.

Start a pilot Back to overview
No PII required to evaluate · FERPA-first by design