Point Lumo, Annoto's AI copilot for education, at any course video and get a draft set of time-tagged questions across Bloom's levels. Review each one, edit what needs it, and publish only what you approve. Published questions behave exactly like hand-written in-video quizzes.
Placing a question takes a minute. Writing a good one takes far longer, so most videos never get quizzed at all.
Questions are tied to the segments they test, across the levels of Bloom's taxonomy.
Every draft is editable and every item can be deleted before anyone sees it.
A quiz stays invisible to students until a human approves and publishes it.
Lumo follows strict privacy and compliance standards.
For how quizzes play, grade and sync, see In-video quizzes →
Lumo's draft appears beside the video as a list of questions, each with its timestamp, question type and Bloom's level. Open any item to edit the stem, answers, points and feedback, or remove it. Approve what you want, then publish.
The pedagogical win is coverage: questions on every video, not only the two or three you had time to build.
Generated questions are ordinary Annoto interactions once published. Everything you can do with a hand-written quiz applies, and the drafting step adds a few things of its own.
Lumo works from what the video actually covers and proposes questions tied to the segments they test, each with a timestamp placed after the explanation.
Each draft is labelled with its cognitive level, from remembering to creating, so you can balance recall checks with application, analysis and evaluation.
True/False, multiple choice, multiple answer and fill-in-the-blank style items, plus open reflection prompts for the levels a fixed-answer question cannot reach.
Rewrite a stem, replace a weak distractor, move the timestamp, change the points, add feedback in your own words, or delete an item that misses the point.
Mandatory answer, retries with a set number of attempts, re-watch, time limit, shuffled answers, Answer Reflection, popup or side panel, and pause on show.
A question is not visible to learners until it is published, and only approved items publish. Edits after publishing sync without re-uploading the video.
Activities can be reused or bulk-imported across videos, courses and instructors, so a reviewed quiz survives from one term to the next.
Answers feed Annoto's analytics, and scores sync to the LMS gradebook through LTI 1.3 exactly as they do for hand-built quizzes.
Auto-generation matters most where the volume of video outruns the time available to write for it.
Years of recorded lectures rarely get quizzed by hand. Start with one high-enrolment video: generate, review the draft against the checklist, publish a handful of questions, and look at the results after two weeks. Then expand video by video. Once written, questions survive across terms, so by the second run the quiz layer is mostly maintenance.
The weekly rhythm becomes: record the lecture, ask Lumo for a draft, spend minutes reviewing instead of an hour writing, and publish a low-stakes formative set before class. Item statistics after the week show which concept to reopen in the live session, and a question everyone missed usually points at the minutes just before it.
Designers supporting many instructors, in a university or a corporate learning team, use drafting to make question coverage consistent across a programme rather than dependent on each teacher's spare time. A shared review checklist and a periodic audit keep quality even, and reviewed activities are reused across courses and instructors.
Before exams, a generated set published as practice, with retries and feedback on, turns the course videos into a place to find weak spots while there is still time to fix them. Because Lumo drafts across Bloom's levels, practice reaches explanation and application, not only recall, and practice sets can stay out of the gradebook.
Three steps, all in the player, with the instructor's approval as the gate between the draft and the learners.
Open the video and ask Lumo for a draft. It proposes a set of questions built from what the video covers, each with a timestamp after the relevant segment, a question type, a Bloom's level and suggested feedback. Nothing is visible to learners at this stage.
Work through the list with a fixed checklist. Rewrite stems, swap distractors, move timestamps, set points, and write feedback in your own words. Choose the settings for each item: mandatory or not, retries, re-watch, time limit, shuffled answers, popup or side panel. Delete anything that misses the point.
Publish the approved items. Learners answer in the player, results feed Annoto's analytics, and scores sync to the LMS gradebook when the set is graded. Per-question statistics show which items work and which need revision, so the quiz improves with every cohort, and edits after publishing sync automatically.
Reviewing a draft is fastest with three fixed questions per item. Anything that fails the first is rewritten or deleted, because a question that needs unseen material teaches students to distrust the quiz.
The question should be answerable by someone who watched carefully and read nothing else. If it depends on outside reading, it belongs in a different assessment.
A good distractor is what a learner with a common misconception would pick. Obviously wrong options make the item a giveaway; replace them with the mistakes you actually see.
Place the question after the explanation ends: close enough that the learner can still reconstruct the reasoning, far enough that answering requires recall rather than echo.
Published questions produce the same data as any Annoto quiz: every answer feeds the analytics, with correct-answer rates per question, completion, and scores that sync to the gradebook. The productive stance is editorial. A question everyone gets right may be doing no work, or confirming a concept is secure. A question half the cohort misses points at the minutes just before it: was the explanation rushed, was a term used before it was defined?
Because each item carries a Bloom's level, the data also shows where understanding thins out. If recall items score well and application items do not, the video explained the idea but never showed it in use. Instructors who review item statistics at the end of a term, revise the weakest segment or the weakest question, and regenerate where the video changed, end up improving the teaching asset itself, not only the quiz around it.
Average first-attempt correct rate for the published set on one video.
A drop at the higher levels shows where the video explains but does not demonstrate.
Activity on one video's quiz set, as reported in Annoto's analytics.
Generated questions inherit everything the in-video quiz layer already does inside Canvas, Moodle, Blackboard, Brightspace and Open edX, and Lumo is built for education-grade privacy and compliance.
Published sets behave like any graded Annoto activity. Scores sync to the LMS gradebook through LTI 1.3, with no export step, and practice sets can stay out of the gradebook entirely. For grading policy and sync details, see In-Video Quizzes.
Annoto runs as a layer over Kaltura, Panopto, Wistia, YouTube and Vimeo. Questions appear as a popup or in the side panel, pause the video if you choose, and can be edited after publishing with changes syncing automatically.
Lumo streamlines content creation, grading and personalised insights while keeping educators in the driver's seat, and is built with strict adherence to privacy and compliance standards. Quiz drafting is one of the things it does. Learn about Lumo AI
“The Annoto Quizzes were easy to set up and it provided students with immediate feedback.”
Book a 20-minute demo. Bring a video, see the draft, review it live, and watch the published questions land in your LMS with grades syncing. One course, one reviewer, the checklist shared up front.