From statements to a decision
The dashboard a manager actually opens.
Most learning dashboards bury managers in charts. This one leads with the four things they came for: who's done, how they scored, who needs a nudge, how long they've got. Then it does what a completion report can't. It reads the raw xAPI, finds the exposure hiding in the answers, and drafts the follow-up.
Assignment
Customer Data Handling
Revenue Operations · managed by Dana Okafor · due Fri, Jun 19
Completion
7/12
58% of the team
Avg. check score
81%
of those who finished
Need coaching
6
failed, stalled, or unstarted
Due in
3 days
Fri, Jun 19
Team roster
Tap a row with a ⚑ to read what they wrote
| Person | Status | Score |
|---|---|---|
PR Priya Raman⚑ AE | Completed | Passed · 100% |
MB Marcus Bell⚑ AE | Completed | Failed · 67% |
SN Sofia Nguyen SDR | Completed | Passed · 100% |
JP Jordan Park⚑ Ops Analyst | Completed | Failed · 33% |
AK Aisha Khan AE | Completed | Passed · 100% |
TB Tom Becker SDR | Completed | Passed · 67% |
EV Elena Vasquez Ops Analyst | Completed | Passed · 100% |
CD Chris Donnelly AE | In progress | — |
MF Maya Foster SDR | In progress | — |
RP Raj Patel AE | In progress | — |
WH Will Hartley AE | Not started | — |
GL Grace Liu Ops Analyst | Not started | — |
This is the whole picture a manager gets today: status, scores, and the open-text answers. The completion rate looks fine, the risk is buried in the ⚑ answers. That's the gap the next step closes.
AI analysis on the xAPI
Turn the statements into a decision.
The scores and the free-text answers get read together, so the exposure patterns surface as something a manager can act on in two minutes. The plan and nudges below are that analysis, ready to go — and the question box underneath puts a live model on the same data for anything you want to ask it.
Follow-up training plan
Gaps in the data → who to coach, on what, by when.
Coaching nudges, ready for Slack
A team digest and per-person DMs, drafted, not preachy.
Ask about this team
Example answersWith a model connected, you can interrogate the roster in plain English. A taste of what that looks like:
“Who should I talk to first, and why?”
Jordan Park. Lowest score on the team (33%) and the reflection names an active habit — a weekly full-table CSV export for the board deck — that an audit would flag today. Marcus Bell (67%) is second: he forwards customer email to personal Gmail and isn’t sure what counts as PII.
“Is this a re-train-everyone situation?”
No. Completion is healthy (7 of 12) and most scores pass. The risk is concentrated in two named, fixable habits — that’s two 15-minute coaching conversations, not a program re-run.
About this demo
The team is fictional, but the data is shaped exactly like aggregated LRS output: one assignment, a roster of learners, per-person launched / progressed / answered / passed / completed statements, quiz scores, and the free-text answers people actually typed. The completion rate looks healthy. The risk is hiding in three of those answers. That's the whole point.
Because this roster never changes, the plan and Slack drafts are the analysis pre-computed — the live model is saved for the question box, where it answers whatever you ask about the same data. In a real engagement it reads fresh xAPI statements posted to a client LRS, and the Slack messages go out through a webhook. The draft-and-approve flow stays exactly as shown.