Assessment Surveys
Build structured workforce AI-readiness surveys with thematic sections and rich question types, segment responses by member tags, watch a Microsoft-Forms-style results page (per-question charts, time-to-complete, and per-person responses, or run it anonymous), synthesize responses into skill gaps, and export for analysis.
What it is
Surveys group questions into thematic sections (e.g. AI familiarity, workflow, sentiment, barriers, readiness). Question types: single-select, multi-select with "select up to X", scale (comfort/confidence), free text, and conditional "Other".
Use
- Build a survey in Admin → Assessment Surveys: add sections, then questions of any type, or let Prism auto-generate questions. Choose whether the survey is anonymous (respondent names and emails are never recorded with results) or named.
- Publish through a governance flow: draft → review → approved → published.
- Learners take it one question at a time with a progress bar; responses feed Prism synthesis into a skill-gap vector.
- The Stats tab shows summary metrics (sent, responses, response rate, average total and active time, days open) plus a card per question with a chart matched to its type, donut for short option lists, bars for long lists and multi-select, a score histogram for scales, and a keyword cloud for free text. A named survey also shows each person's individual answers; an anonymous survey shows a locked state instead.
- Export responses to CSV with the respondent's member tags for Excel/Power BI.
Setup
Create a survey, add sections and questions, send it to members (or share a link), and publish it. Responses are captured per member; synthesis runs after responses arrive.
Assign tags to members first (Admin → Members) so responses can be segmented by tag in cross-tabs.
Security
- Draft surveys cannot accept responses; only published surveys are open.
- Recipient gating: only invited members may respond (unless a survey was never sent to anyone).
- Server-side validation enforces completeness, "select up to X" limits, and required "Other" text, the client is never trusted.
- Anonymous surveys never expose or record respondent identity in results, per-person responses, cross-tabs, synthesis labels, or CSV export; all are gated server-side, not just hidden in the UI.
- Every survey and response is org-scoped; cross-tenant reads and writes are rejected.