Real briefs. Blind review. A score you would hire on.
Trailhead runs on work that matters to someone. Employers supply the problem, review the result, and see a portfolio that predicts job performance better than a credential does. Reviewing takes about an hour per learner, twice a program. Here is exactly what that hour is.
The rubric, as the form scores it
Work Quality
25%
AI Tool Utilization
20%
Problem-Solving
20%
Communication
15%
Responsible AI Practices
10%
Collaboration
10%
7.0 of 10 is the bar for the AI-Ready mark. The score makes a learner eligible; the facilitator issues it.
Five things, one page. We help you write it.
A brief is not a syllabus. It is the problem as you would hand it to a new hire, with the quality bar stated. Learners work it through the Trailhead Loop with Scout, and ship one artifact to their portfolio.
Writing one takes an hour with us on the call. Two per program is typical.
- A real problem
- Something your team would otherwise do, defer, or pay for. Not a simulation and not a case study.
- A real audience
- Who reads the result and what they do with it. That is the standard the work is judged against.
- A deliverable
- One artifact, named: a spec audit, a recommendation memo, a dashboard, a decision ladder, a playbook page.
- Constraints
- Budget, data you can share, data you cannot, the deadline. Constraints are where the judgment shows.
- A quality bar
- What a 7 looks like on your desk. Scout drafts the criteria from your brief, each linked to the sentence it came from, and you approve them. The reviewer scores the rubric.
Scout checks the work against your bar first.
Between panels, learners submit their brief work and Scout checks it against the criteria you approved. Your time goes to work worth your judgment. Scout never scores, and never counts toward the certificate.
You approve the criteria
Scout drafts them from your brief. You edit and approve them, or we do it with you on the call. Learners cannot submit until a person has approved the bar.
Scout checks every submission
A written response, attachments, a reflection and an optional "How I used AI" note. Not ready yet means one question per criterion for the learner, never the answer and never a score.
You receive only work that passed
A short email tells you there is work waiting, with no learner details in it. You read it blind, marked only "Passed Trailhead’s check".
Return it with feedback
If it is not ready for your desk, send it back with a note. That is formative: the learner reads it, revises and resubmits. Scoring on the rubric happens at the panels.
For school cohorts, the teacher reviews everything too. Sending work to an outside employer starts once the district's data-sharing agreement covers it; until then staff review it. Sponsor groups inside your own organization receive passed work directly.
About an hour per learner, and it is the hour that counts.
Panels run twice a program. Before the first one you spend two hours with us calibrating on exemplar artifacts, so two reviewers reading the same work land on the same score.
You read the work blind
A pseudonym, the submitted snapshot, nothing else. Identity never reaches a reviewer, so the score is about the work.
The process trace sits beside it
The rule the learner invented, their answer to the counter-case, the rules they taught and then fixed, their transfer attempts. The version history authenticates the work. We use no AI-detection tools anywhere.
Ten minutes of oral defense
You pick one artifact and ask them to explain the choices in it. The form records what you heard. It is the fastest honest signal there is.
Two raters, then adjudication
Every capstone is scored by two trained raters against anchor exemplars. When the two straddle 7.0 or land more than a level apart, a third adjudicates. We report chance-corrected agreement, never percent agreement.
You never see:
Student names or emailsCoach transcriptsScout’s notes on the workPer-learner telemetryAnything outside your windowTrailhead for Teams
Sponsor the work from inside your own organization.
On the Teams track the employer can be you. A sponsor group is a handful of people in your workspace who post the briefs, approve the criteria Scout checks against, and score the results: the same team as the learners, another department, or a client you invite. Same blind review, same trace, same oral defense, filed under the group. Sponsors see the artifact scores and a cohort transfer summary for their own briefs, and nothing per person.
Your team sponsors itself
Put the same people in a sponsor group. The manager who learns also assigns.
Another department sponsors
Ops writes the brief; the analysts ship the audit. The review goes back to Ops.
A client sponsors
Invite them through the external portal: scoped, time-boxed, audited.
Two briefs, two panels, a year.
Intake and onboarding. Two to three partners signed by month 3.
First brief assigned to a cohort.
First deliverables submitted.
First portfolio review panel.
Second brief cycle; second panel; micro-internship placements.
Paid micro-internships, ten to twenty hours.
The review corpus is what you remember a learner by: what got a 7 and why. The natural next step is a short, paid piece of real work, scoped with us and the school. The pilot targets three or more placements in its first year. Nobody is obliged to hire; everybody gets to see the work first.