Financial services landscape: banking, markets, payments, insurance, wealth & asset management. AI vs ML vs statistical models in a financial context. Financial data types: transactional, behavioral, market, reference, and unstructured documents. AI/ML lifecycle: framing → data → modeling → deployment → monitoring. Risks & limitations: explainability, data quality, and black-box concerns.

Data sources & pipelines: core banking, trading, CRM, KYC/AML, market data vendors. Architectures: data lakes, warehouses, lakehouse patterns; batch vs streaming pipelines. Data quality, lineage, metadata management for regulated environments. Feature engineering & feature stores for risk, fraud, personalization. Infra controls: access, encryption at rest/in transit; on-prem/cloud/hybrid patterns.

Credit decisioning, segmentation, next-best-offer, churn prediction, and collections optimization. Trade surveillance, anomaly detection, and NLP for research and news summarization. Document understanding for KYC/onboarding/loan docs and workflow automation. GenAI co-pilots for relationship managers/traders with approved knowledge sources. Scenario selection: choosing the right approach for a given business problem.

Credit risk concepts: PD, LGD, EAD; scorecards vs ML trade-offs. Fraud & AML: anomaly detection, graph methods, false positive reduction, alert triage. Stress testing & scenario analysis (high level). Operational risk & conduct risk signals (conceptual). Monitoring: drift detection and sensitivity vs precision threshold tuning.

Wealth: robo-advisory concepts, risk profiling, personalization for advisors. Research: NLP over filings and disclosures; GenAI summarization with human oversight. Insurance: underwriting support from structured/unstructured inputs. Claims triage and automation; fraud detection in claims. Workflow integration in advisor/underwriter portals and client reporting.

Governance lifecycle: inventory → validation → approval → deployment → review → retirement. Documentation standards: model dev docs, validation reports, change logs. Model risk categories: data, implementation, conceptual, and use issues. Auditability & traceability; backtesting, challenger models, and overrides. Independent validation and benchmarking as part of sound MRM practice.

Fairness & non-discrimination in lending, pricing, and customer treatment. Explainability: local vs global explanations; practical requirements for decisions. Human oversight: escalation, human-in-the-loop, and appeals processes. Customer transparency about AI use and automation limits. Internal policies and governance committees for higher-risk use cases.

GenAI in finance: co-pilots, document automation, conversational enterprise interfaces. AI + digital assets/DeFi/CBDCs (high level): analytics, surveillance, and risk signals. Real-time payments and embedded finance: AI for instant risk and fraud decisions. Operating models: talent, standards, and centers of excellence. Scenario questions: future-proof roadmaps aligned to strategy and risk appetite.