Skip to demonstrator
RILayer

Controlled demonstrator

Youth Transition · YT-WM-001

Exit demo

YT-WM-001 · Critical-function proof

A machine says “do not progress.” RILayer makes the decision governable.

See the complete governed change first. The detailed governance trail and selective Discernment functions are available underneath.

Adult 18+ · Synthetic · Not live

01 · Machine recommendation

DO NOT PROGRESS

A fictional screening system has recommended that Chloe should not progress.

02 · Evidence is checked

NOT READY FOR AN ADVERSE FINAL DECISION

The available evidence does not yet justify accepting the machine recommendation as the final answer.

03 · Evidence change

CHRONOLOGY RECONCILED

New clarification resolves an important evidence conflict.

04 · Human action

PROCEED TO STRUCTURED INTERVIEW

A named human reviewer makes the decision and records the reason. Chloe separately retains the choice to accept or decline the interview.

Three engineered boundaries to notice
  • The screening system’s recommendation must be reviewed — it is not the final decision.
  • Employment gap must not be converted into an unsupported reliability judgement.
  • Autism or an adjustment request must not be treated as a TRAUMA signal or selection evidence.

A simple way to follow this example

  1. 01Machine recommends
  2. 02Evidence is checked
  3. 03Risks and boundaries are checked
  4. 04An accountable human decides

The nine-step governance trail remains below for detailed traceability and audit review.

CASE ORIENTATION

Before you start

Chloe is a fictional adult early-career candidate. A simulated recruitment screening system has recommended DO NOT PROGRESS.

RILayer does not replace that recommendation with another automatic decision. It helps an accountable person check the evidence, uncertainty, risks and decision boundaries before deciding what should happen next.

Fictional demonstration · No real candidate · No live AI · Final authority remains human

As you move through the demonstrator, notice three things:

  • The machine recommendation stays visible, but it does not become the final decision.
  • Evidence and assumptions are separated before action is taken.
  • An accountable human makes and records the final decision.

60-SECOND GOVERNANCE PATH

Experience the governance change

Controlled click-through · existing synthetic Chloe case · not a live recruitment system · not an open-ended AI sandbox.

SIMULATED MACHINE RECOMMENDATION

DO NOT PROGRESS

Original simulated input retained in the case record for audit continuity.

View original simulated machine input

61/100

Not a RILayer score or a validated measure of candidate suitability.

DUAL-LANE ARCHITECTURE

Two people. Two decisions. One controlled handoff.

RILayer keeps Chloe's private reflection separate from the employer's progression decision.

CHLOE'S DECISION

Candidate lane — Chloe's decision

Does this opportunity fit me?

Chloe decides whether the next step fits her direction, priorities and practical situation. Her private reasoning stays in her protected space.

Candidate-Controlled Discernment

MINIMUM AUTHORISED HANDOFF

Status only — not private reasoning

Only the minimum authorised status needed for the next governed step crosses between the two lanes.

EMPLOYER DECISION

Employer lane — progression decision

Does the authorised evidence justify progression?

The employer decides whether the approved evidence supports offering a structured interview. That decision does not decide whether Chloe accepts the interview.

Case Governance + accountable human action

Chloe's private reflection does not become employer selection evidence.

RILayer governs both decision boundaries without collapsing them into one decision.

Where this demo sits in RILayerAdult early-career transition · 18+ synthetic use case

Discover happened before this case. This demonstrator shows Discern only.

Adult boundary: ‘Youth Transition’ in YT-WM-001 refers to adult early-career transition. The current controlled prototype is not designed or authorised for use with minors.

DISCOVER

BEFORE RILAYER

Completed before governed RILayer

DISCERN

YOU ARE HERE

CURRENT DEMONSTRATED PHASE

DEVELOP

LATER

Wider governed pathway

DEDICATE

LATER

Wider governed pathway

What this demonstrator is showing

DISCERN — CURRENT DEMONSTRATED PHASE

DISCERN is the RILayer lifecycle phase demonstrated here. Within this case, two governed lanes are visible.

Candidate lane

Candidate-Controlled Discernment

Selective Discernment functions help Chloe examine her own decision.

Minimum authorised status only

Employer lane

Employer Case Governance

Case Governance controls examine the machine input, authorised evidence, Decision Readiness, Governance Risk, boundaries and accountable human decision.

Case Governance is not an additional RILayer model and is not a candidate assessment.

Canonical governance terms — with plain-English translations

Decision Readiness

Plain English: Do we have enough legitimate evidence to make the next decision properly?

Decision Readiness concerns the evidence, clarity, boundaries and authority available for the next decision. It is not a score of Chloe.

Governance Risk

Plain English: What could go wrong in the decision process?

Governance Risk belongs to the evidence, interpretation, system, process, boundaries or decision pathway. It is never a risk score about Chloe.

Candidate-Controlled Discernment

Plain English: Chloe examines her own decision.

RILayer supports the decision process without choosing a career direction for her.

Minimum authorised handoff

Plain English: Only the necessary status crosses the boundary.

Private reasoning does not become employer selection evidence.

Employer Decision Governance

Plain English: Is the employer's next decision properly evidenced and authorised?

The employer-side process governs the machine input and progression decision without taking Chloe's own decision away from her.

Purpose-bound non-routing

Plain English: If a function is not relevant, RILayer does not use it or invent data to activate it.

Non-collection is a governance control.

ALIGN — Coherence check

Plain English: Do the relevant parts now make sense together?

ALIGN must not become an employability score, suitability score or automated permission.

Discover: Discover is the pre-governed orientation space. It may be supported by SCOPE (opens in a new tab), self-reflection, a career adviser, mentor or another trusted environment. Only a sufficiently clear human question and authorised relevant context need enter governed RILayer. Discover does not score, diagnose, prescribe, activate RILayer models or make the decision.

Discernment RoutingSynthetic routing · not live model delivery

What must be checked before anyone acts?

Only the checks relevant to this decision are opened. Technical RILayer names are shown for traceability. The human decision stays with the accountable person.

Chloe's question

“Should I continue towards finance and business operations, or widen my direction because progress has been slower than expected?”

PRE-RILAYER DISCOVER

How the question became clear enough to bring into RILayer

No score · no diagnosis · no recommendation
SCOPE reflection (opens in a new tab)Career adviserMentor / trusted personChloe's own reflection
What each Discover source contributed
  • SCOPE reflection (opens in a new tab): Optional private self-reflection environment
  • Career adviser: Guided transition conversation
  • Mentor / trusted person: Practical perspective and reflection
  • Chloe's own reflection: Candidate-originated concern after seeing the vacancy

Discover is the pre-governed orientation space. It may be supported by SCOPE (opens in a new tab), self-reflection, a career adviser, mentor or another trusted environment. Only a sufficiently clear human question and authorised relevant context need enter governed RILayer. Discover does not score, diagnose, prescribe, activate RILayer models or make the decision.

Used in this walkthrough · 5 functions

What must be checked before anyone acts?

Plain-English purpose comes first. Technical RILayer names are shown for traceability, with routing detail available on demand.

Learning & progressionStabilityIndependenceFair opportunity

Direction

RILayer function: DREAM

ROUTING SIGNAL PRESENT

Direction remains meaningful to test: finance / business operations.

Why this function is here

Chloe is questioning whether slower progress means she should leave her chosen finance / business-operations direction.

Boundary: RILayer does not recommend a career direction for Chloe.

Pause before interpreting

RILayer function: CALM

MANDATORY GOVERNANCE CHECK

Mandatory CALM check pending before evidence interpretation.

Why this function is here

If the simulated ATS returns DO NOT PROGRESS, Chloe may be tempted to treat the machine output as a verdict about her direction.

Boundary: Non-clinical reflective pacing only; no distress score or diagnosis.

What matters

RILayer function: VALUE

ROUTING SIGNAL PRESENT

The opportunity still fits important priorities, but practical feasibility needs clarification.

Why this function is here

Learning, progression, stability, independence and fair opportunity matter to Chloe's next step.

Boundary: VALUE does not morally rank Chloe's priorities.

Practical feasibility

RILayer function: MONEY

ROUTING SIGNAL PRESENT

Practical / financial feasibility still needs evidence.

Why this function is here

Salary and location are known; travel pattern, cost and support are not yet fully known.

Boundary: RILayer does not recommend what salary Chloe should accept, provide budgeting advice or make financial decisions for her.

Coherence check

RILayer function: ALIGN

COHERENCE CHECK PENDING

Not evaluated yet because required evidence is unresolved.

Why this function is here

Direction, priorities, reflective availability, practical feasibility and decision evidence must cohere before the next human-owned step.

Boundary: ALIGN must never become a suitability score or automated permission.

Purpose-bound non-routing · Considered, but not routed · 3 functions

RILayer does not invent a reason to use every model

TRAUMA · RHYTHM · CARE — no relevant signal is present. Expand to inspect why each function is deliberately not routed.

Non-collection is a control

Safety & safeguarding boundary

RILayer function: TRAUMA

Not routed
Why it is not routed

No trauma or safeguarding signal is present in this synthetic case.

Outcome: No TRAUMA routing in YT-WM-001.

Boundary: Autism or an adjustment request must not be treated as a TRAUMA signal.

Routine / pacing

RILayer function: RHYTHM

Not routed
Why it is not routed

No material routine, energy or pacing pattern requires separate routing for this decision.

Outcome: No RHYTHM routing in YT-WM-001.

Boundary: Do not invent a pattern simply to exercise the model.

Relational / caring load

RILayer function: CARE

Not routed
Why it is not routed

No material caring or relational-load signal requires separate routing for this decision.

Outcome: No CARE routing in YT-WM-001.

Boundary: Do not collect caring information without a relevant purpose.

More about Chloe's synthetic context
  • Chloe wants to continue developing towards finance and business operations.
  • Slower-than-expected progression has caused her to question whether she should stay on that path or broaden into general administration.
  • She values learning and progression, stability, financial independence and fair evidence-based consideration.
  • The trainee role appears relevant, but working-pattern and travel affordability are not fully known.
  • She feels pressure to make progress quickly, creating a legitimate need for reflective pacing before interpreting an adverse machine signal.

Practical questions still open: Required on-site days · Likely travel cost / frequency · Training or travel support · Any timing constraint affecting a start date

Five functions are relevant. Three are not needed.

Next, the employer-side Case Governance flow examines the machine input, evidence and human decision without taking Chloe's decision away from her.

Want to inspect the proof? The full case record shows how the original machine input, evidence change, governed route and human action stay traceable.

The linked full case record is the canonical reference record for YT-WM-001. The current-session decision record is shown inside Stage 9 of the guided walkthrough.

Routing is signal-led and purpose-bound. RILayer uses only relevant functions and does not collect, infer or expose information merely to complete every model.

View full case record

Demonstrator conclusion

What happened in this fictional example?

The original machine recommendation was not accepted automatically. Evidence and assumptions were reviewed, and an evidence conflict was clarified.

The relevant risks and decision boundaries were checked. An accountable human recorded the governed next step. RILayer did not make the employment decision.

This is a fictional proof-of-concept. A real pilot would test the governance process using an organisation’s authorised people, evidence and decision rules.

NEXT VALIDATION PATHWAY

From Demonstrator to Real-World Validation

The Chloe Taylor demonstrator shows how RILayer governs one consequential Youth Transition decision before accountable human action.

PAVRA Youth Transition is the proposed next validation pathway.

SCP Ltd's immediate partner-development focus is the West Midlands, where we are seeking dialogue with suitable stakeholders about the case for, design of and requirements for a potential bounded pilot.

No pilot has yet been commissioned or agreed.

Continue from the proof-of-concept

Explore the PAVRA validation pathway or discuss whether a bounded West Midlands validation conversation is relevant.

Explore PAVRADiscuss West Midlands Validation