The frameworks
Every assessment on shyfts.work is anchored to published frameworks: SFIA 9 for professional skills, the RISE Compass for human-AI ways of working, RECALL for certified micro learnings, and SAPERE for graduate candidate assessment.
RISE Compass — twelve human skills
Each person self-assesses twelve skills at Level 1–5, shown as a spider-web profile. The four skill groups are labels only—not scores, person placements or employment measures. RISE supports development and evidence-led mentor conversation.
1 · Foundational — Recognises the skill and applies it with guidance in a familiar situation.
2 · Practitioner — Applies the skill independently in routine professional work and explains the approach used.
3 · Advanced — Applies the skill in complex or changing work, adapts the approach to context, and can validate or correct AI-supported work from first principles.
4 · Strategic — Improves team practice, workflows, standards, or controls for the skill and helps others build it.
5 · Visionary — Influences wider organisational or professional practice and anticipates how the skill should evolve.
Level 2 requires independent routine application and explaining the approach. Strong AI output alone is not evidence. Original self-levels remain unchanged; mentors may add a separate confirmed view.
Knowledge Application
How do we apply what is known to produce a robust outcome?
- Design — Turn requirements, constraints, user needs, and available technologies into a coherent solution shape.
- Orchestrate — Coordinate people, AI tools, workflows, and resources from intent to a reliable outcome.
- Review — Check work for completeness, correctness, usability, alignment, and fitness for purpose.
Sensemaking & Decision
What is happening, what matters, and what should we decide?
- Reason — Build a sound line of thinking from evidence, context, assumptions, and possible explanations.
- Evaluate — Assess options, outputs, risks, evidence quality, and consequences against explicit criteria.
- Judge — Make a defensible decision under ambiguity, uncertainty, or competing values.
Transactional Execution
How do we execute repeatable work reliably and improve it over time?
- Translate — Turn a work request, rule, decision, or need into clear instructions, specifications, prompts, tickets, or workflows.
- Automate — Convert repeatable work into reliable workflows, scripts, agents, templates, or system routines.
- Monitor — Observe a workflow, identify errors, drift, exceptions, or control failures, and take or trigger corrective action.
Exploration & Creativity
What could be possible, and how might we turn it into something worth testing?
- Explore — Move through uncertainty without closing the question too early; scan possibilities and test unusual combinations.
- Originate — Create a distinctive idea, concept, product direction, story, or solution—not merely select an existing option.
- Plan — Turn creative possibility into a coherent path for learning, testing, or development.
SFIA 9 — 147 skills
Strategy and architecture
31 skills
- Strategy and planning
- Financial and value management
- Security and privacy
- Governance, risk and compliance
- Advice and guidance
Development and implementation
41 skills
- Systems development
- Data and analytics
- User centred design
- Content management
- Computational science
Delivery and operation
31 skills
- Technology management
- Service management
- Security services
- Data and records operations
Relationships and engagement
15 skills
- Stakeholder management
- Sales and bid management
- Marketing
Change and transformation
16 skills
- Change implementation
- Change analysis
- Change planning
People and skills
13 skills
- People management
- Skills management
RECALL — certified micro learnings
RECALL is the certified micro learning layer of shyfts.work. It gives mentees short, regular guided practice that rebuilds core skills and first-principles thinking, so people keep the reasoning that AI would otherwise do for them. Mentors assign RECALL trainings from the catalogue their organisation has approved, and link each one back to case studies, feedback and development goals.
Human-reviewed RISE and SFQ
Programme design—not the learner—is reviewed: RISE at least 20/36 (twelve evidence scores of 1–3), SFQ at least 4/5 and no critical safeguard below 3. Only human-certified micro learnings reach the catalogue.
Assigned in context
Mentors choose from the approved catalogue and assign trainings alongside the case study a mentee is working on.
Tracked to evidence
Completion, mentor feedback and reflection are recorded on the mentee's development path, not in a separate LMS.
SAPERE — graduate assessment, 7 dimensions
SAPERE assesses graduate candidates for agency, discernment and adaptability. Candidates work a real case rather than answering screening questions, so thinking is visible rather than generated. Each dimension is marked 1-5 by a human reviewer, with AI use disclosed and verified alongside the score.
Pillars: Agency · Discernment · Adaptability
Problem framing
AgencyHow clearly does the graduate define the real problem, its scope and what success looks like?
Level 5: Sharp, defensible framing with explicit scope, constraints and success measures.
Inquiry and questioning
AgencyDoes the graduate ask the questions that would change the answer, and identify missing information?
Level 5: Targets the decisive unknowns and states how each answer would change the approach.
Evidence and AI verification
DiscernmentIs evidence sourced, checked and attributed, and is AI-assisted work disclosed, verified and owned?
Level 5: Every material claim is sourced and checked; AI use is disclosed, verified and owned.
Options and trade-offs
DiscernmentAre genuine alternatives generated and weighed against stated criteria?
Level 5: Real alternatives weighed against explicit criteria, with the rejected paths explained.
Socratic response
AdaptabilityUnder challenge, does the graduate reason further rather than capitulate or dig in?
Level 5: Absorbs the challenge, tests it, and updates or defends the position with new reasoning.
Applied solution
AgencyIs the proposed action specific, feasible and fit for the role context?
Level 5: Specific, sequenced and feasible within the stated constraints.
Reflection
AdaptabilityDoes the graduate say what remains uncertain, what they would test, and how their thinking changed?
Level 5: Honest uncertainty, a named test for it, and a clear account of how the thinking changed.