{"stable_id":"8cd32f6eda7ef8fb","slug":"jpmorgan-coin-contract-review","company":"JPMorgan Chase","workflow_name":"COiN agreement extraction","function_code":"credit","pattern_codes":["exception_based_operations","intake_redesign"],"changed_assumption":"Defined attributes can be extracted across annual volume while lawyers handle interpretation exceptions.","evidence_strength":"reported","publication_tier":"showcase","freshness":"current","reviewed_at":"2026-08-23","updated_at":"2026-08-22","source_quality_summary":"1 independent, 1 primary; publication outcomes are reported.","caveat_summary":"Company upper-bound estimate; seconds is machine extraction, not end-to-end disposition.","freshness_caveat":null,"workflow_summary":{"before":[{"actor":"Lawyers and loan officers","action":"Read commercial-credit agreements and manually identify relevant clauses and attributes.","actor_type":"control"},{"actor":"Loan-servicing staff","action":"Use interpreted terms in servicing processes and correct interpretation errors.","actor_type":"human"}],"hinge":"Defined attributes can be extracted across annual volume while lawyers handle interpretation exceptions.","after":[{"actor":"COiN","action":"Uses unsupervised machine learning to extract 150 defined attributes from agreements.","actor_type":"system"},{"actor":"Lawyers and loan officers","action":"Review exceptions and retain legal interpretation and servicing decisions.","actor_type":"system"}],"decision_mode":"shared","decision_marker":"AI handles the default; humans own exceptions"},"before":[{"order":1,"actor":"Lawyers and loan officers","action":"Read commercial-credit agreements and manually identify relevant clauses and attributes.","handoff_to":"Loan-servicing data entry","control":"Approximately 12,000 annual agreements; company estimated up to 360,000 aggregate hours."},{"order":2,"actor":"Loan-servicing staff","action":"Use interpreted terms in servicing processes and correct interpretation errors.","handoff_to":"Servicing system","control":"Human legal interpretation and servicing controls."}],"after":[{"order":1,"actor":"COiN","action":"Uses unsupervised machine learning to extract 150 defined attributes from agreements.","handoff_to":"Legal/loan exception review","control":"Defined attribute schema; machine extraction reported in seconds, not end-to-end disposition."},{"order":2,"actor":"Lawyers and loan officers","action":"Review exceptions and retain legal interpretation and servicing decisions.","handoff_to":"Servicing system","control":"Unknown or low-confidence clauses require human review; no public error-rate threshold."}],"decision_rights":"Humans own legal interpretation and downstream servicing.","exception_path":"Unknown clauses or low-confidence extraction require human review.","removed_work":["Manual extraction of defined clauses"],"outcomes":[{"metric":"Processing effort","baseline":"As many as 360,000 hours/year","result":"150 attributes from 12,000 agreements extracted in seconds","period":"Initial 2016 implementation","scale":"About 12,000 commercial-credit agreements annually","attribution_caveat":"Company upper-bound estimate; seconds is machine extraction, not end-to-end disposition.","evidence_label":"reported"}],"executive_lesson":"Constrain AI to a defined extraction schema.","anti_pattern":"Do not imply complete lawyer replacement.","questions_for_leaders":["Where is the operating threshold set and who can override it?","What measured result would trigger rollback or retraining?","Which residual decisions must remain human-owned?"],"collections":["human-still-decides","queue-eliminated","regulated-autonomy"],"bundle_version":"1.0.0","bundle_fingerprint":"sha256:c23c6cc2b88153d008ea8fda928f632ce0011fc2d4c5036672a16e5d895bab93","canonical_url":"https://brianletort.ai/transformations/jpmorgan-coin-contract-review"}