{"stable_id":"a3327c8f8ced4e18","slug":"danske-bank-ai-fraud-monitoring","company":"Danske Bank","workflow_name":"Machine-learning and deep-learning transaction fraud scoring","function_code":"financial_crime","pattern_codes":["continuous_decisioning"],"changed_assumption":"That handcrafted rules are the primary way to identify digital payment fraud.","evidence_strength":"reported","publication_tier":"showcase","freshness":"watch","reviewed_at":"2026-08-23","updated_at":"2026-08-22","source_quality_summary":"1 independent, 1 vendor; publication outcomes are reported.","caveat_summary":"Outcome figures originate from Danske executives and Teradata materials; independent Forbes reporting quotes the same executive, not a separate audit.","freshness_caveat":"Watch status: verify the cited source and deployment condition before reusing this case.","workflow_summary":{"before":[{"actor":"Rules engine","action":"Apply human-authored fraud rules to transactions.","actor_type":"control"},{"actor":"Fraud investigator","action":"Review up to 1,200 alerts per day, most of which are false positives.","actor_type":"control"}],"hinge":"That handcrafted rules are the primary way to identify digital payment fraud.","after":[{"actor":"ML and deep-learning engine","action":"Score millions of transactions in real time using learned patterns and latent features.","actor_type":"system"},{"actor":"Fraud investigator","action":"Investigate the smaller, higher-yield alert queue and decide customer or enforcement action.","actor_type":"system"}],"decision_mode":"moved","decision_marker":"Selection moves from a fixed rule to the model"},"before":[{"order":1,"actor":"Rules engine","action":"Apply human-authored fraud rules to transactions.","handoff_to":"Fraud investigators","control":"Static rule thresholds"},{"order":2,"actor":"Fraud investigator","action":"Review up to 1,200 alerts per day, most of which are false positives.","handoff_to":"Customer or enforcement action","control":"Human investigation"}],"after":[{"order":1,"actor":"ML and deep-learning engine","action":"Score millions of transactions in real time using learned patterns and latent features.","handoff_to":"Fraud investigators","control":"Sub-300ms scoring target"},{"order":2,"actor":"Fraud investigator","action":"Investigate the smaller, higher-yield alert queue and decide customer or enforcement action.","handoff_to":"Customer and case systems","control":"Human adverse-action authority"}],"decision_rights":"The model prioritizes and scores; investigators determine whether activity is fraudulent and what action to take.","exception_path":"Investigators review alerts, use model explanations, and escalate or clear cases; the public sources do not disclose automated blocking rights.","removed_work":["A large share of false-positive alert review","Exclusive dependence on handcrafted fraud rules"],"outcomes":[{"metric":"False-positive alerts and true-positive detection","baseline":"Up to 1,200 false positives per day; 99.5% of investigated cases not fraud; about 40% fraud detection","result":"Vendor case study reports 60% fewer false positives and 50% more true positives","period":"Initial 2017 production deployment","scale":"Millions of online banking transactions scored in real time","attribution_caveat":"Outcome figures originate from Danske executives and Teradata materials; independent Forbes reporting quotes the same executive, not a separate audit.","evidence_label":"reported"}],"executive_lesson":"The transformation metric is investigator yield, not model accuracy in isolation; a lower false-positive queue changes how scarce human judgment is allocated.","anti_pattern":"Automating account action from an opaque score or presenting vendor-reported uplift as independently audited.","questions_for_leaders":["What is the residual false-negative risk?","Which actions require human confirmation?"],"collections":[],"bundle_version":"1.0.0","bundle_fingerprint":"sha256:c23c6cc2b88153d008ea8fda928f632ce0011fc2d4c5036672a16e5d895bab93","canonical_url":"https://brianletort.ai/transformations/danske-bank-ai-fraud-monitoring"}