Questions institutions ask us

A career development platform your institution deploys for its students, with a command centre your team runs it from. One system, two front doors, one set of numbers underneath. Students get a measured picture of where they stand and a sequence of things to do about it. Your institution gets cohort-level readiness, a placement pipeline that maintains itself, and curriculum drift measured against what employers are actually hiring for.

No. The assessment is one early step. What follows it is the product: career matching, gap analysis against a target role, learning roadmaps that are verified rather than self-ticked, live job matching, interview practice with a coaching report, and resume work. SkillDrift is also not a resume builder.

It is a weighted blend of nine measured components, computed in the backend: skill assessment, experience and evidence, situational scenarios, career alignment, voice, academic record, written reflections, supporting documents and engagement quality. The weighting is our own calibration and we do not publish it, but we walk procurement committees through the full model under NDA, and your institution can tune the weighting on top of it. What matters for trust is the property, not the numbers: it is arithmetic rather than a language model’s judgement, so running it twice on the same evidence gives the same number.

Yes, and this is deliberate. Reassessment is adaptive and bidirectional. It targets a student’s weakest prior dimensions and spot-checks areas they were previously strong in. If they fail a retest in an area they had been credited for, the credit is taken back. A score that can only rise is a score nobody should make decisions on.

The missing component scores zero across its full weight and the weight is not redistributed. Skipping the voice assessment genuinely lowers the score. The report names what was skipped, links the student to it, and the score recalculates when it is completed. A separate data quality modifier scales the result, so a thin attempt does not produce a confidently wrong number.

Twenty to thirty minutes, once. One question per screen, saved continuously, and resumable across devices, so a student can start it on a laptop between lectures and finish it later without losing anything. It is designed to be taken once rather than repeatedly, and retakes are granted by your team rather than self-served. What follows it is where the time actually goes: the career report, the gap analysis, the learning roadmap and the interview practice.

Single sign-on over SAML 2.0 or OIDC, including Google and Microsoft. LMS integration over LTI 1.3 into Moodle, Canvas and Blackboard. Student information and ERP integration with SAP Campus and Oracle PeopleSoft. Placement CRM sync with Salesforce and HubSpot. Live market openings arrive over an API feed. You can also bulk-invite by email or CSV and skip integration entirely for a pilot.

Each institution runs on its own subdomain with its own isolated database. Every request resolves to one institution before it touches any data. Access is role-based and enforced at the API rather than hidden in the interface, with super admin, admin and manager tiers plus custom roles you define. There is a complete audit trail, including an automatic record of every mutating administrative request. Data is encrypted in transit and at rest, and residency and retention are configurable.

Yes to FERPA and GDPR alignment, and yes to HECVAT. Ask us and we will send the HECVAT workbook, our GDPR data processing documentation and our written FERPA position.

AI generates situational questions around a student’s own background, analyses voice, writes the career report narrative, matches careers, builds learning roadmaps, scores job fit, runs interview practice and writes coaching reports. It does not compute the Career Readiness Score, which is arithmetic. It does not write the reason a score moved, which comes from the component difference. It is not in the path of the cohort heatmaps, so any cell traces back to real responses. In the Command Center it receives anonymised aggregates only and writes interpretation, and no figure on that page is model-generated. Every AI call is metered and attributed by activity, model and student, with a full ledger.

Every institutional screen is computed from what students actually did. Nobody keys anything in to keep a dashboard current. The two exceptions are deliberate: your team manages the Interviewing and Placed stages of the pipeline, because those are the only two the system cannot observe, and both have full move history and a one-click return to the automatic stage.

Try it first. The pilot is free: up to 25 students, eight weeks, and a measured result at the end. There is nothing to buy and no purchase order to raise before you know whether it works on your own students. You get their readiness map inside two weeks and a reassessment at week eight, so you can see whether the number moved. Pick a deliberate mix rather than your best students, because the point is to see whether the score reads them the way you already do. If it works, the platform is one price per student per year, quoted as a single flat number, with no per-module pricing and no surprise line items.