{"stable_id":"953d873c30c6de37","slug":"hsbc-dynamic-risk-assessment","company":"HSBC","workflow_name":"Dynamic Risk Assessment transaction monitoring","function_code":"financial_crime","pattern_codes":["exception_based_operations","creator_to_judge"],"changed_assumption":"Transaction-monitoring teams do not need a large rules-generated queue to preserve risk coverage; machine-learning risk detection can narrow the queue while investigators retain consequential review.","evidence_strength":"mixed","publication_tier":"showcase","freshness":"current","reviewed_at":"2026-08-23","updated_at":"2026-08-22","source_quality_summary":"2 primary, 1 vendor; publication outcomes are verified and reported.","caveat_summary":"Company-reported comparison without independent outcome validation; 'suspicious activity' and 'financial crime found' are not defined as confirmed offenses.","freshness_caveat":null,"workflow_summary":{"before":[{"actor":"Rules-based transaction-monitoring system","action":"Screen transactions against predefined parameters and generate alerts.","actor_type":"control"},{"actor":"Financial-crime investigators","action":"Manually review flagged transactions, including a high volume of false positives.","actor_type":"control"}],"hinge":"Transaction-monitoring teams do not need a large rules-generated queue to preserve risk coverage.","after":[{"actor":"Dynamic Risk Assessment","action":"Score transactions and identify suspicious activity using an AI model trained on HSBC data.","actor_type":"system"},{"actor":"Financial-crime investigators","action":"Review a smaller, more risk-concentrated alert queue.","actor_type":"system"}],"decision_mode":"shared","decision_marker":"Human moves from creator to judge"},"before":[{"order":1,"actor":"Rules-based transaction-monitoring system","action":"Screen transactions against predefined parameters and generate alerts.","handoff_to":"Financial-crime investigators","control":"Rules encode the monitoring thresholds."},{"order":2,"actor":"Financial-crime investigators","action":"Manually review flagged transactions, including a high volume of false positives.","handoff_to":"Customer contact or further investigation","control":"Investigators determine whether flagged activity warrants action."}],"after":[{"order":1,"actor":"Dynamic Risk Assessment","action":"Score transactions and identify suspicious activity using an AI model trained on HSBC data.","handoff_to":"Financial-crime investigators","control":"HSBC trains and assesses the model under its responsible-AI practices."},{"order":2,"actor":"Financial-crime investigators","action":"Review a smaller, more risk-concentrated alert queue.","handoff_to":"Customer contact, escalation, or law-enforcement reporting","control":"Humans retain investigation and escalation decisions."}],"decision_rights":"The AI system prioritizes risk; HSBC investigators retain case review, customer-contact, escalation, and reporting decisions.","exception_path":"Suspicious or uncertain activity is routed to investigators for manual review and escalation.","removed_work":["A reported 60% of the prior alert volume","Manual review of many rules-generated false leads","Several weeks of batch-processing delay"],"outcomes":[{"metric":"Transaction-monitoring alert volume","baseline":"Rules-based monitoring during the 12 months before go-live.","result":"HSBC reported 60% fewer alerts while identifying two to four times as much suspicious activity.","period":"12 months after go-live versus the preceding 12 months, with seasonality adjustment and the first post-go-live month removed.","scale":"More than 1.2 billion transactions screened monthly in the 2023 HSBC-authored account; absolute alert counts were not disclosed.","attribution_caveat":"Company-reported comparison without independent outcome validation; 'suspicious activity' and 'financial crime found' are not defined as confirmed offenses.","evidence_label":"reported"}],"executive_lesson":"The defensible value is not automation of the compliance decision; it is a smaller, higher-yield investigation queue with humans retaining consequential case decisions.","anti_pattern":"Treating alert reduction as proof of crime prevention or allowing the model to close consequential cases without investigator review.","questions_for_leaders":["Which queue metric will be compared before and after deployment: alerts, cases, or confirmed suspicious activity?","Which decisions remain with investigators?","How will model drift and missed-risk rates be monitored?"],"collections":["regulated-autonomy"],"bundle_version":"1.0.0","bundle_fingerprint":"sha256:c23c6cc2b88153d008ea8fda928f632ce0011fc2d4c5036672a16e5d895bab93","canonical_url":"https://brianletort.ai/transformations/hsbc-dynamic-risk-assessment"}