01Key Responsibilities
- Review application, bureau, banking, financial and fraud information for eligibility and affordability.
- Understand model/score outputs, key drivers and policy rules rather than treating automated decisions as unquestionable.
- Perform manual review for referred, borderline or exception cases with documented rationale.
- Verify customer data and investigate inconsistencies or potential manipulation.
- Apply credit policy, delegated authority and fair-customer treatment requirements consistently.
- Escalate policy exceptions and model concerns to credit risk owners.
- Track approval, decline, override and performance outcomes to improve policy and models.
- Never use unapproved AI to make or explain credit decisions.
- AI-enabled ways of working:
- Use approved GenAI copilots, enterprise search/RAG, analytics and workflow agents to reduce repetitive work and improve decision preparation.
- Keep prompts, outputs and automated actions within approved data-access, confidentiality and retention rules.
- Correct inaccurate summaries, classifications or recommendations before they become customer, operational or system records.
- Feed recurring AI errors, knowledge gaps and process friction into product, documentation and control improvement.
- Human accountability: AI is a copilot, not the accountable decision-maker.
- The role holder remains responsible for judgement, quality, privacy, security and appropriate escalation.
- Credit, fraud blocking, legal/regulatory, financial advice, safety-critical, clinical, employment and other high-impact decisions must follow delegated authority and required human review.
- Success Measures:
- Decision turnaround improves
- Override quality and consistency improve
- Bad-rate/fraud outcomes stay within appetite
- Decision explanations are documented and defensible .
