{"stable_id":"d31443b5cf716855","slug":"sbi-ai-cheque-stp","company":"State Bank of India","workflow_name":"AI straight-through processing for low-value cheques","function_code":"engineering","pattern_codes":["threshold_as_control","straight_through_processing"],"changed_assumption":"A bounded low-value tier can move to touchless processing while assurance shifts to sampling.","evidence_strength":"reported","publication_tier":"showcase","freshness":"current","reviewed_at":"2026-08-23","updated_at":"2026-08-22","source_quality_summary":"2 independent; publication outcomes are reported.","caveat_summary":"Executive disclosure through independent financial outlets; no error rate.","freshness_caveat":null,"workflow_summary":{"before":[{"actor":"Cheque-processing staff","action":"Reads low-value cheques and verifies mandatory fields and compliance requirements.","actor_type":"human"},{"actor":"Operations staff","action":"Clears or routes the cheque after manual verification.","actor_type":"human"}],"hinge":"A bounded low-value tier can move to touchless processing while assurance shifts to sampling.","after":[{"actor":"Vision-LLM/STP system","action":"Reads and validates eligible cheques up to ₹10,000 and passes qualifying items without routine touch.","actor_type":"system"},{"actor":"Control Risk Unit","action":"Samples AI-passed cheques, identifies misses, and determines whether retraining is needed.","actor_type":"system"}],"decision_mode":"shared","decision_marker":"Human sets the threshold; AI decides each instance"},"before":[{"order":1,"actor":"Cheque-processing staff","action":"Reads low-value cheques and verifies mandatory fields and compliance requirements.","handoff_to":"Payment-processing queue","control":"Human review was the stated prior operating mode; sources do not disclose baseline time or error rate."},{"order":2,"actor":"Operations staff","action":"Clears or routes the cheque after manual verification.","handoff_to":"Core payment system","control":"Bank payment controls and value limits apply."}],"after":[{"order":1,"actor":"Vision-LLM/STP system","action":"Reads and validates eligible cheques up to ₹10,000 and passes qualifying items without routine touch.","handoff_to":"Core payment system or exception queue","control":"Bank-set ₹10,000 scope; approximately 25% of cheque volume."},{"order":2,"actor":"Control Risk Unit","action":"Samples AI-passed cheques, identifies misses, and determines whether retraining is needed.","handoff_to":"Model and operations owners","control":"Human sampled assurance; above-threshold and exception handling details are not quantified publicly."}],"decision_rights":"AI/STP passes eligible cheques; bank sets threshold; CRU audits samples.","exception_path":"Above-threshold or uncertain cheques remain human-reviewed.","removed_work":["Per-cheque human review below ₹10,000"],"outcomes":[{"metric":"Touchless share","baseline":"Human cheque verification","result":"About 25% of cheque volume processed with practically no human involvement","period":"FY2026","scale":"SBI cheque operations","attribution_caveat":"Executive disclosure through independent financial outlets; no error rate.","evidence_label":"reported"}],"executive_lesson":"The design is the monetary boundary plus sampled control lane.","anti_pattern":"Do not infer accuracy or savings from touchless volume.","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":["queue-eliminated"],"bundle_version":"1.0.0","bundle_fingerprint":"sha256:c23c6cc2b88153d008ea8fda928f632ce0011fc2d4c5036672a16e5d895bab93","canonical_url":"https://brianletort.ai/transformations/sbi-ai-cheque-stp"}