aimsdog
Create
SystemAIS-004Financial services

Consumer Credit Scorecard

Score unsecured consumer credit applications and return a probability of default with the four principal reasons for the score.

Model tier
Tier 1, material
Status
In operation
State
Deployed
Owner
Louis Wu
Record completeness
56

100, less 12 for each open blocker and 4 for each warning. It measures the record, not the performance of the AI System.

The system is in operation while its record stands below the floor of 70. That contradiction is raised as blocker STATE-01.

Blockers
3

Lifecycle

Where the system stands, and the act that moves it on. Every move is written to the audit trail under the acting persona.

Registered

The system exists as a record. Design work has not started.

In development

Design and development are under way. Nothing is in use outside the project.

Approved for operation

The release decision is taken and the record supports it. The system is not yet in use, so the post-market duties have not started.

In operation

In use outside the project. The post-market duties apply from this point.

Under restriction

In use, but with a limitation placed on it after a signal or a finding.

Suspended

Use is stopped while a matter is settled.

Retired

Withdrawn from use. Records are held for the retention period.

Move to
Rationale

Identity

What the system is, where it is placed, and how often it is reviewed.

Description
Tier 1 model. Independent validation is annual. A challenger model runs in shadow on every scored application.
Markets
European Union
Sectoral class
Tier 1, material
FDA position
Not assessed
EU AI Act tier
Not assessed
QMS reference
MRM-REG-0142
Deployed configuration
Scorecard 7.1
Placed on market
Yes
Internal reviews
12 months, next due 2027-02-24 Open AI Reviews

Linked records

Navigate from the AI System to every governed record.

Chronology

Every dated act on this AI System and its records, in order.

  1. 2026-07-25SignalDrift: Population stability index risingPSI 0.19 on the applicant mix against a 0.25 action threshold. Monitoring.
  2. 2026-02-24ReviewAIS-004Periodic review completed
  3. 2026-02-20Releaseb3c1907Release v7.1Refit on 2025 book, add affordability features
  4. 2023-09-11SystemAIS-004AI System registered

Governance gates

A gate is held until it is proven. Each card states what is already in place and, plainly, the single next act that would move it. A gate can hold on one unmet item while everything else is done.

Development authorisation

Authorises development against a stated intended purpose.

Held
Next act

No AI Use Case is recorded, so there is nothing to authorise development against. Record a use case

Then 1 further requirement.

In place · 1
  • The intended purpose is not stated on the AI System.
Responsible · Independent Reviewer
Model approval

Approves an immutable AI Model Version for use.

Held
Next act

No AI Model Version exists, so there is nothing to approve. Record a model version

Then 2 further requirements.

Responsible · Independent Reviewer
System approval for operation

Approves the complete AI System for operation in a market.

Held
Next act

No AI Risk is recorded, so the risk position is unstated. Record a risk

Then 2 further requirements.

  • CONF-01 No deployed AI Model Version
  • FIN-01 Independent validation is not current
Responsible · Independent Reviewer
Change authorisation

Authorises a change inside the Controlled Evolution boundaries.

Held
Next act

No Controlled Evolution boundary is recorded for this AI System. Record the boundary

Then 2 further requirements.

Responsible · Quality and Regulatory
Resumption

Resumes operation after restriction or suspension.

Held
Next act

The AI System is not restricted or suspended, so there is nothing to resume. Open the AI System

Then 1 further requirement.

Responsible · Independent Reviewer

Posture summary

Compliance posture narrative

A written summary of the computed findings and the recommended order of work. The findings themselves are deterministic. The narrative never states that the AI System is compliant.

Open findings

3 blockers, 2 warnings. Readiness is 100, less 12 for each open blocker and 4 for each open warning. It measures how much of the record is unfinished, not how well the AI System performs.

  • BlockerCONF-01No deployed AI Model Version

    The AI System is approved for operation but no model version is marked deployed.

    Effect
    The deployed configuration cannot be evidenced.
    Record the deployed version
  • BlockerFIN-01Independent validation is not current

    A material model is validated by a function independent of the developer, at least annually.

    Effect
    Blocks continued use of the model in production decisions.
    Plan the validation review
  • BlockerSTATE-01The AI System is in operation while its record is unfinished

    2 other blockers are open against this AI System, which puts readiness at 68. A system in use is expected to stand at 70 or above with no blocker open.

    Effect
    The state declared on the record is not supported by the records behind it. Either finish the record, or move the system back to a state it can support.
    Open gaps and readiness
  • WarningCE-03No Controlled Evolution declared

    Permitted change is undefined, so every change needs a new authorisation.

    Effect
    No PCCP can be generated.
    Declare Controlled Evolution
  • WarningFIN-02No fairness risk is recorded

    A model that decides access to a financial service carries a discrimination hazard, whether or not it is realised.

    Effect
    Weakens the risk file and the Article 10 data governance position.
    Record the fairness risk

Monitoring

Signals observed on the deployed configuration.

MetricObservedThresholdStateObserved on

Framework mappings

A mapping records where a requirement is addressed. It does not assert compliance.

FrameworkClauseHow it is addressedEvidence
EU AI ActArticle 14, Human oversightAdverse decisions above a stated exposure are routed to a credit officer.Not linked

Audit trail

  • 2023-09-11Louis WuRegistered AI System
  • 2026-02-24Lilith IyapoCompleted independent validation