Baseline diagnostic
One defined cohort establishes the starting readiness signal and trust-risk pattern.
Methodology
How STEP turns AI use into workplace judgment evidence.
STEP targets the gap universities cannot solve with tool training alone: the judgment line managers expect from new hires. Students use AI, test it, explain it, and deliver work that can be reviewed against a clear standard.
University delivery model
A complete 6-week pathway from baseline to results review.
STEP Campus gives universities the whole loop: baseline, structured practice, certification assessment, Individual Development Plans, fresh reassessment, and a leadership review on what graduates can already handle.
Six-week learning arc
Students practise the STEP behaviours through realistic AI-era workplace situations.
Fresh simulation evidence
The credential decision uses assessment forms students cannot pass by memorising practice tasks.
University action plan
Faculty, careers, and leadership receive group evidence, IDPs, and next steps.
Where STEP fits
AI literacy is the baseline. STEP tests use at work.
As higher education moves AI skills into the core curriculum, universities still need to know whether graduates can apply those skills with judgment. STEP sits after or alongside AI-literacy modules as the capstone evidence layer: the place where students show they can question, judge, and own AI-assisted work.
Students learn AI foundations
Universities teach what AI is, how it works, and how students can use it within their disciplines.
STEP creates workplace pressure
Students face unclear briefs, realistic AI errors, people pressure, and limited time.
The cohort report shows the gap
Faculty and careers teams see where AI competency becomes trusted work, and where it still needs practice.
Policy context
Built for the next question after Singapore's 2027 AI competency direction.
Singapore's higher-education direction points toward baseline AI competencies, discipline-specific application, responsible guardrails, and humans staying at the helm. STEP is aligned with that direction by testing whether students can use AI in work a manager would trust.
AI competency becomes expected
Universities teach common foundations and field-specific use. STEP tests what students do with that knowledge at work.
Responsible use must be visible
Students must show source checks, privacy judgment, clear AI-use explanation, and where AI should not replace human responsibility.
Judgment stays human-owned
Students evaluate, challenge, and take responsibility for AI-assisted work instead of accepting output too quickly.
6-week STEP Campus program
The full pathway from AI literacy to workplace trust evidence.
The program is designed to sit after or alongside a university's AI-literacy modules. It gives students repeated practice before the credential decision, and gives the university evidence it can use.
00
Baseline Diagnostic
Before Week 1
Establish the starting signal
Students complete an initial diagnostic so the university can see baseline readiness, trust-risk flags, and the priority gaps for the cohort.
01
Week 1 · Structure
Ambiguous work
Turn a vague request into a bounded task
Students practise clarifying the audience, decision, source boundaries, assumptions, and AI-use plan before rushing into output.
02
Week 2 · AI Boundaries
Safe commissioning
Use AI without giving away the work
Students learn what context AI needs, what should stay out, and how to keep human judgment accountable for the final work.
03
Week 3 · Test
Claim audit
Catch polished but unsafe output
Students separate supplied facts, AI-generated claims, assumptions, and unknowns before a confident draft becomes manager rework.
04
Week 4 · Engage
People and privacy
Handle constraints without hiding behind the tool
Students practise people judgment, privacy boundaries, escalation discipline, and concise disclosure under pressure.
05
Week 5 · Perform
Delivery under pressure
Ship useful work with visible judgment
Students deliver manager-ready work samples: briefs, AI audits, client-safe updates, handoff notes, and revised recommendations.
06
Week 6 · Assessment + Results
Credential decision
Assess, report, and decide the next move
Students complete fresh simulations without coaching. The output is a credential decision, Individual Development Plan, cohort report, targeted practice path, and fresh reassessment requirement where needed.
Inside the capstone
Students face the moments that make or break trust.
The scenario is structured enough to score and realistic enough to reveal judgment. Students must use AI, verify it, explain it, and ship usable work.
01
Phase 1 · The Brief
10–15 min
Receive an ambiguous brief
A simulated manager sends an incomplete brief with a tight timeline. The student decides what to clarify, what to assume, and how to structure the work. This phase assesses Structure.
02
Phase 2 · Research and Draft
60–90 min
Use AI, then test the result
The student uses AI to research, analyse, and draft. The AI output contains errors that look correct: weak sources, outdated data, confident wrong claims, or biased framing. This phase assesses Test.
03
Phase 3 · People Pressure
20–30 min
Respond to people pressure
A colleague or client changes the brief, questions the approach, or raises an AI-use concern. The student must respond without hiding behind the tool. This phase assesses Engage.
04
Phase 4 · Revision and Delivery
30–45 min
Deliver usable work under pressure
The student revises and delivers under time pressure. They cannot redo everything, so the assessment is whether they make sound choices and ship usable work. This phase assesses Perform.
05
Phase 5 · Reflection
10–15 min
Explain the judgment behind the work
The student explains what AI did, what they verified, where they made hard choices, and what they would do differently. This adds a self-awareness layer to the four scored dimensions.
How scoring works
The rubric turns decisions into a readable record.
- Structured decisions capture how the learner frames the brief, verifies claims, and handles AI risk.
- Human-reviewed moments capture judgment a short answer would miss.
- Scenario variants keep the task realistic and reduce answer-sharing.
- Shared standards give faculty, careers teams, and learners one language for the result.
- Human-at-the-helm evidence checks that AI supports judgment instead of replacing it.
- AI available, not automatic means students must frame the task and inspect sources before AI help becomes useful.
Each dimension is described on a three-level scale: Developing (1), Competent (2), Proficient (3). Certification requires Level 2 or above across all four dimensions. The purpose is to make judgment visible, not to turn the experience into a black-box test.
Rubric preview
Faculty can inspect what the credential means.
Each score is tied to observable work, not confidence, polish, or personality.
Dimension
Evidence signal
Competent performance
Structure
Work brief, assumptions, source boundaries
Frames the decision, names missing context, and defines what AI should and should not do.
Test
AI audit trail and claim checks
Catches false claims that look correct, weak evidence, and unsupported certainty before delivery.
Engage
Manager note, disclosure, pressure response
Explains judgment clearly and protects trust when the work changes or remains uncertain.
Perform
Final recommendation or decision note
Ships usable work that a line manager can act on without hidden cleanup.
Assessment quality
Built for academic review, not black-box scoring.
The method can be explained to faculty, careers teams, boards, and quality committees before it becomes part of a graduate readiness review.
Start with the claim
Every mission starts with the claim, the proof needed, and the scenario moment that can reveal it.
Use work products
Students produce briefs, audits, disclosures, and recommendations instead of recall answers.
Anchor the judgment
Scoring guides and example answers help reviewers handle borderline cases consistently.
Say what is proven
STEP shows simulation performance. Broader accreditation, hiring, or norm claims require institution-specific proof.
Make the standard inspectable
Universities can review the rubric, scoring records, and group results through faculty or academic panel processes.
Keep humans accountable
Students must show what they trusted, checked, changed, explained, and kept under human judgment.
The output
One run, multiple university uses.
Students get a credential and scorecard. The university gets a group results review. Faculty, careers teams, and leadership use the same work samples for different decisions.
- For learners — a verified credential and plain-language scorecard.
- For institutions — group patterns for advising, curriculum, and graduate readiness reporting.
- For faculty — a clearer view of which AI-era capabilities students have actually demonstrated.
- For leadership — a concise summary on group strengths, gaps, and next actions.
The scorecard travels with the learner. The cohort report stays useful to the institution. Public verification confirms the credential without exposing underlying student work.