Docebo reports completion, progress and assessment results, but has no visibility into attention within a piece of content. Emotuit adds content-level engagement data — dwell, focus, tab-switching and drop-off position — through a sub-50KB script, alongside existing reporting rather than replacing it.
What Docebo already gives you
Docebo Analytics provides course completion, enrolment, learner progress and a custom report builder. Docebo Learn Data adds a warehouse layer for larger deployments.
What it does not
Comprehensive at the enrolment and completion layer, silent below it. You can see that 62% completed a course and nothing about which of its twelve lessons cost you the other 38%.
This is not a criticism of Docebo specifically — it is true of essentially every LMS on the market. Completion and navigation data is what these systems were built to record, and they record it well. Attention is a different measurement problem, solved in the content layer rather than the platform layer.
How Emotuit fits
Inject via the Custom CSS/JS feature in Admin Menu → Settings → Advanced Settings, or through a Docebo Page widget. Content positions map to training material IDs.
The SDK is under 50KB, adds one script tag, and requires no migration or infrastructure change. It runs alongside Docebo's own reporting rather than replacing it.
Details
Where the data lands
Engagement scores map to Docebo learning object IDs and can be pushed back as custom fields.
Works with
Docebo-hosted content, SCORM and xAPI packages, and embedded video.
What you get
| Metric | Detail |
|---|---|
| Content-level engagement | A score for every section, not one per course |
| Drop-off position | Where sessions actually end |
| Focus-adjusted dwell | Real attention, discounting background tabs |
| Tab-switch rate | The strongest single predictor of disengagement |
| Replay behaviour | Where learners went back — a confusion signal |
| Cohort comparison | A/B test content versions on real attention |
Deployment options
Emotuit runs in two configurations, and for LMS deployments in the EU the choice matters.
Signals uses behavioural telemetry only — no camera, no biometric data, no emotion inference. It sits outside the EU AI Act Article 5(1)(f) prohibition on inferring emotions in workplace and education contexts, which makes it the default recommendation for corporate L&D and education.
Learn adds a client-side facial expression layer. It should not be deployed to EU employees or students, and is intended for market research, UX and media testing contexts.
Frequently asked questions
Does this replace Docebo's built-in reporting?
Does it need a webcam?
How long does integration take?
Will it slow the platform down?
See it on your own Docebo content
We'll instrument one module and show you the engagement data before you commit to anything.