Survey Builder

The builder opens on a three-pane canvas: a left sidebar listing question types and template blocks, a center editor that renders the live questionnaire as the respondent will see it, and a right inspector for properties, validation, and logic. A new survey starts either from scratch or by cloning a template from the library, which seeds the canvas with question groups, default copy, and branching rules already wired up. Authors drag question types onto the canvas — SingleChoice, MultiChoice, Likert, NPS, Matrix, Ranking, ShortText, LongText, Number, Date, DateRange, ImageUpload, GeoPoint — and each card exposes its own settings panel for choices, scale anchors, validation, and required flags.

Conditional branching is configured through a visual rule builder that composes AND/OR trees over previous answers; the canvas highlights downstream questions that become hidden or required based on the active rule, so authors can preview the flow without leaving the editor. Translations are managed inline through a language switcher in the toolbar: each question card exposes a translation drawer with side-by-side source and target strings, status badges (missing, machine-suggested, reviewed), and bulk import/export via CSV for translation agencies. Templates are first-class objects with a category taxonomy (member satisfaction, post-event CSAT, rule consultation, regulatory consent), versioning, and ownership at tenant, federation, or global level, so a CEP-authored consultation template can be promoted to the shared library and cloned by any member nation.

Every save produces a draft revision with autosave; publishing locks the schema and assigns a stable question_id so analytics remain stable across edits. Authors can preview the form on desktop, tablet, and phone breakpoints, and a built-in validator surfaces issues such as orphan branches, untranslated strings, or unreachable questions before publication.