HR
The onboarding that ends when the PDF is sent
The new hire gets six documents on day one and a “any questions, just ask”. Nobody knows what they actually read — not even them, three weeks later.
Understanding measured
Platform · Quizzes
“Everyone read it” is not an answer.
When a policy changes, someone sends the PDF and writes “please read carefully”. After that there is no information at all: not who opened it, not who understood it, not which part was confusing. A quiz closes that gap starting from the document your company has already approved — with nobody writing questions from scratch.
See how it worksBORN FROM THE DOCUMENT
point at the policy or procedure and the questions come from there, not from the internet
PEOPLE AND GROUPS
bulk assignment from the company directory, with a deadline and a note
75% PASS MARK
a default your company changes, along with time, attempts and shuffling
ERRORS PER QUESTION
results show where the team got it wrong, not just who passed
Where this shows up in your day
The training is over. So what do you actually know?
You know who was in the room. You do not know who understood. And when something goes wrong months later, the only evidence the subject was covered is an attendance list that proves nothing about comprehension.
Operations
The procedure that changed and nobody noticed
The standard was updated, the notice went out in the group chat, and the team carried on the old way. The mistake only surfaces at the next audit, once it has already cost something.
Who still does not know
Quality
The question everyone gets wrong
If half the team misses the same question, the problem is not the team: it is a standard that reads ambiguously or training that never covered it. Without results per question, that stays invisible.
The blind spot shows
How it works
From document to answer, with you reviewing in between.
A quiz is not a form somebody types up over an afternoon. It grows from content the company already has, passes through your review, and becomes an assignment with a deadline.
01
Choose the source
Say what the assessment should measure and point at the documents behind it — the policy, the procedure, the manual. The questions come from that content, not from the model's general knowledge.
02
Set the size
Three to twenty questions, two to six alternatives each, and the difficulty: easy for straight facts, medium for application, hard for scenario analysis. Every question has exactly one correct answer.
03
Review before releasing
The draft opens in the editor: reorder, rewrite a stem, fix an alternative, change which one is correct. Whoever knows the subject has the final word — the AI prepares, it does not publish on its own.
04
Assign with a deadline
Pick people from the company directory or whole groups — a group becomes one assignment per person. Set the deadline and leave a short note explaining why. Anyone already assigned is not assigned twice.
05
See what came back
Per person: who answered, what they scored, whether they passed. Per question: where the team gets it wrong. That second number usually says more about the material than about the people.
Live proof
The standard changed on Tuesday. By Friday you know who understood.
The scene that matters is not somebody answering questions: it is the loop closing — approved content, an assigned assessment, and a number that says where clarity is still missing.
Starting point
The customer-data policy has been revised and published in the company's knowledge. Thirty-two people in support need to know what changed.
What happens
- 01The training lead creates the quiz pointing at the revised policy as the source, asks for eight medium-difficulty questions, and gets a draft ready to review.
- 02She fixes the wording of two questions, replaces an alternative that read ambiguously, and assigns it to the whole Support group, due Friday, with a one-line note.
- 03Each person answers once, with questions and alternatives shuffled. The correct answers only appear after submitting — and only if the company turned that option on.
Result
By Friday there is a number instead of an assumption: who answered, who reached the pass mark and — most useful of all — that the question about retention periods was missed by half the team.
What this scene does not promise
- This is not certification and not proof of learning. It is an internal assessment with a record of who answered and what they scored — nothing here issues a certificate with external standing.
- Questions are multiple choice, with exactly one correct alternative. There are no essay answers and no free-text grading.
- The assessment is generated from the documents you point at, which is why it needs human review before it reaches the team. The quality of the source is the ceiling on the quality of the question.
To evaluate carefully
What you can set before it reaches the team.
Select a topic to see its capabilities, conditions, and limits.
- Generated from your documents
- Questions grow from the documents you point at as the source. Material that has already been reviewed and approved internally is what becomes an assessment.
- Size and difficulty
- Three to twenty questions, two to six alternatives each, and three levels: straight facts, application, or scenario analysis.
- Review is required
- The draft opens in the editor so you can reorder, rewrite stems and fix alternatives. Nothing reaches the team without someone having the final word.
- Pass mark
- The default is 75% and your company changes it per assessment, along with whether someone who fell short may retake it.
- Time and attempts
- A time limit per attempt and a maximum number of attempts are optional: without them there is no clock and no ceiling.
- Shuffling
- Questions and alternatives come out in a different order for each person, by default. It is what keeps answers from being traded in the corridor.
- The answer key at the right time
- While someone is answering, which alternative is correct never leaves the assessment. Showing the answers afterwards is the company's choice.
- Assign to groups
- A whole group can be assigned at once; each person gets their own assignment, with a shared deadline and note. Duplicates are ignored.
- Results in two readings
- Per person (who answered, score, passed or not) and aggregated per question — which shows where the material, not the team, needs work.
Getting it running
Start with the document everyone ought to know.
You do not need a training programme to begin: you need one current document and an honest question about who really knows it.
- 01Whoever runs training
Points at the current policy or procedure and generates the draft assessment.
- 02Whoever knows the subject
Reviews stems and alternatives in the editor before any assignment.
- 03Whoever runs training
Sets the pass mark, time, attempts and whether answers are shown at the end.
- 04Administrator
Assigns to people or company groups, with a deadline and a note.
- 05Whoever follows up
Reads the per-person result and the per-question aggregate to decide the next step.
Included
- Question generation from the documents you point at.
- A review editor with question reordering and alternative fixing.
- Bulk assignment to people and groups, with a deadline and a note.
- Per-person results and an aggregated view of accuracy per question.
Your company decides
- Pass mark, time limit, number of attempts and whether retaking is allowed.
- Whether correct answers appear afterwards and whether results go out by email.
- Which documents serve as the source for each assessment.
- Who is assigned, with what deadline, and in which hub or workspace the assessment lives.
Frequently asked
What usually holds the decision back.
Start with one document
Pick the standard everyone ought to know.
Information security, customer service, onboarding for new hires. Generate the assessment from it, assign it to the team, and find out — with a number — what they actually understood.