{"stable_id":"6ff380bc84cd5443","slug":"fortune500-genai-support","company":"Anonymous Fortune 500 software company","workflow_name":"Generative AI agent assistance","function_code":"software_engineering","pattern_codes":["creator_to_judge"],"changed_assumption":"AI can diffuse expert practice, with value concentrated among novices.","evidence_strength":"verified","publication_tier":"showcase","freshness":"current","reviewed_at":"2026-08-23","updated_at":"2026-08-22","source_quality_summary":"2 independent; publication outcomes are verified.","caveat_summary":"Anonymous company; staggered observational rollout; heterogeneous effects.","freshness_caveat":null,"workflow_summary":{"before":[{"actor":"Customer-support agent","action":"Reads the incoming chat, searches available knowledge, and drafts a response.","actor_type":"ai"},{"actor":"Support agent or manager","action":"Resolves the issue or escalates when managerial intervention is needed.","actor_type":"ai"}],"hinge":"AI can diffuse expert practice, with value concentrated among novices.","after":[{"actor":"Generative conversational assistant","action":"Observes the live chat and proposes response text based on patterns from prior interactions.","actor_type":"system"},{"actor":"Support agent","action":"Accepts, edits, or ignores the suggestion and owns the sent response and escalation.","actor_type":"ai"}],"decision_mode":"shared","decision_marker":"Human moves from creator to judge"},"before":[{"order":1,"actor":"Customer-support agent","action":"Reads the incoming chat, searches available knowledge, and drafts a response.","handoff_to":"Customer or manager escalation","control":"Control group had no generative assistant; agents could handle concurrent chats."},{"order":2,"actor":"Support agent or manager","action":"Resolves the issue or escalates when managerial intervention is needed.","handoff_to":"Customer","control":"Human resolution and escalation authority."}],"after":[{"order":1,"actor":"Generative conversational assistant","action":"Observes the live chat and proposes response text based on patterns from prior interactions.","handoff_to":"Support agent","control":"Suggestion only; the working paper does not disclose the company or vendor."},{"order":2,"actor":"Support agent","action":"Accepts, edits, or ignores the suggestion and owns the sent response and escalation.","handoff_to":"Customer or manager","control":"Human acceptance control; effects must be monitored by agent skill because top-agent gains were minimal."}],"decision_rights":"Agent owns every sent response; managers own escalation.","exception_path":"Incorrect suggestions are ignored or escalated.","removed_work":["Some manual search and drafting"],"outcomes":[{"metric":"Issues resolved per hour","baseline":"Agents without access","result":"About 14% average increase; 35% for novice/low-skill agents","period":"2020–2021","scale":"5,179 agents","attribution_caveat":"Anonymous company; staggered observational rollout; heterogeneous effects.","evidence_label":"verified"}],"executive_lesson":"Skill distribution is an operating metric.","anti_pattern":"Do not apply the average uniformly.","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":[],"bundle_version":"1.0.0","bundle_fingerprint":"sha256:c23c6cc2b88153d008ea8fda928f632ce0011fc2d4c5036672a16e5d895bab93","canonical_url":"https://brianletort.ai/transformations/fortune500-genai-support"}