01 · Organize
Build the organization before you measure it.
01
Structure
Create and edit the organization tree.
02
People
Import users and maintain reporting lines.
03
Scope
Apply roles, managers and unit boundaries.
04
Measure
Run hierarchical surveys with a locked core.
Organization context stays attached from administration through to results.
Structure turns one long response list into group-level insight.
A manager sees the branch they are responsible for, while the organization keeps the overall view.
The tree is built once and reused by every survey after it.
Set up the organization
Set up the organization and the people in it.
Import users from a CSV and the unit paths in the file build the tree as it goes — a row with /Engineering/Backend creates both units if neither exists.
Editable reporting lines: a main manager from the person's team, plus extra project managers.
People and units can change without rebuilding the survey.
People and scope
Link people to the same branches they work in.
Organization units on one side, people and managers on the other — Sciences with its science managers, Chemistry with its staff, Chemistry 101 with its class participants.
That correspondence controls participation, manager scope and how results roll up.
Each person sees the organization at the right level. Access follows the active organization and the assigned unit.
An Organization Manager works across the tree, an Org Unit Manager stays inside the subtree they manage, a User sees their own unit.
Hierarchical surveys
One shared survey, and every unit asks what it needs to ask.
The organization has a tree of units, and each unit has a manager. Inside one shared hierarchical survey, each manager can add questions for their own unit and the units below it.
Questions travel down
A question added at Sciences is answered by Chemistry and Biology too. Every unit below the one that added it inherits the question, without anyone rebuilding a survey.
Never sideways
A question added at Chemistry stays in Chemistry — Biology has no access to it. Departments can ask what only they need to ask without adding noise to anyone else's survey.
The core stays locked
Locked entities cannot be modified or deleted. The same question means the same thing in every department, so results stay comparable across the organization.
Fine-grained AI control
Guide the generation when the survey needs more precision.
You set the target audience — who will respond. The objective — what you want to know. The plan — which sections and how many questions. Then the questions themselves: generate, edit and approve.
Save the work, return later, revise any step.
Alternatives
Generate alternatives for a single question and keep the one that fits. Alternative answer choices work the same way.
Focused editing
Reword one question while every other question in the survey stays untouched.
Question types
Pick a type and answer it, exactly as a respondent would.
Four groups — Selection, Rating, Textbox and Forms. Every type below is live: nothing you enter is sent anywhere.
Choose a question type
Pick a type to see what respondents will get.
Selection
Rating
Textbox
Forms
Selection · Multiple Choice
Which of these did you use this term? Select all that apply.
Try it — nothing you enter here is sent anywhere.
Templates, question-bank items, rating presets and instrument outlines are reusable, so a question you have already worded well can be used again instead of rewritten.
Administration
One administration flow from invitation to analysis.
01
Invitation sent
Choose the organization, a unit or a public link.
02
Survey started
People respond from a phone, tablet or desktop.
03
Survey completed
Track final responses before the deadline.
04
Ready for analysis
Results open when the privacy rules allow.
Choose a managed audience or an anonymous public link.
Assign units and invite by email, or open a public link that needs no account. Answers arrive in the same place.
Each run has an opening time and a closing deadline, and a repeat-response guard keeps one person from answering twice.
Respondent experience
Participation works wherever the respondent is.
Open
An assigned survey or a public link.
Answer
From a phone, tablet or desktop.
Review
Check responses before finalizing.
Submit
Completion becomes ready for analysis.
Mobile access also supports quizzes in class.
Survey administration
Track participation before the response window closes.
Each run has an opening time and a closing deadline, and you can follow where completion is still low.
Sent, accepted, responses and awaiting — with awaiting becoming missed once the run closes.
Choose anonymity or identity to match the purpose. An anonymous public link hides who answered.
An identified run is what quizzes and paired studies need, because they compare the same person across rounds.
Privacy threshold
Small groups stay private until enough people respond.
The threshold is applied before group results, free text or comparisons are shown. Five responses, in this example.
Too few responses
A group below the threshold has no score to show. The row says so plainly — too few (needs 5) — rather than showing a number nobody should see.
Hidden to protect a smaller group
A related group may stay hidden as well, when showing it would reveal the smaller group's result by subtraction. That row reads hidden to protect a smaller group.
Results and reporting
Answers become patterns you can read.
Counts and averages per question, scores by unit and by level.
Hierarchical results
Compare units while privacy rules remain in force.
Drill into a unit and compare the questions it shares with the rest of the organization.
Every unit keeps its place in the tree, you compare inside the scope you're allowed to see, and the threshold from the previous section stays in force while you do it.
Gains instruments
Matched responses show what changed and how confident we are.
Before and after are two rounds of the same paired study, matched person by person.
Perceived and demonstrated gains are reported side by side, never merged into one number. Open-text answers are grouped into themes.
Confidence intervals, effect sizes and unit comparisons.
AI summary, inside a paired study
Inside a paired study, an AI summary writes up the numbers shown. It invents nothing and never claims that one thing caused another.
AI comment analysis, per survey
A per-survey switch decides whether AI may read open text at all. Turn it off and respondents' comments are never sent to the AI for that survey.
Every score, count and percentage in the illustrations above is an example, chosen to show how the product behaves. None of it is Survey Teams data or customer data.
Measure change without losing the organization in between.
See it on your own organization.
Build the tree, lock the core, send it out.