ATC Decision-Support Agent ROI Packet

{"ok":true,"packet":{"finance_packet":"atc-decision-support-roi-proof-packet","vertical":{"id":"atc-decision-support","name":"ATC Decision-Support Agent","market":"ATC simulation, traffic-flow management, controller training, and future human-in-the-loop conflict-resolution support","buyer_promise":"Generates ranked conflict-resolution and traffic-flow options in simulator/offline settings with full provenance, replay, and safety-case evidence."},"proof":{"vertical_id":"atc-decision-support","label":"ATC simulator decision support","cost_exposure_label":"Annual training, delay-analysis, and workload-study exposure","annual_cost_exposure_usd":55000000,"verified_improvement_pct":4.5,"capture_rate_pct":45,"confidence_discount_pct":68,"annual_platform_cost_usd":600000,"pilot_cost_usd":150000,"total_first_year_cost_usd":750000,"conservative_annual_gross_savings_usd":757350,"conservative_annual_net_savings_usd":7350,"payback_months":11.9,"first_year_roi_multiple":0.01,"measurement_window_days":120,"proof_metric":"simulator throughput, conflict-resolution quality, workload score","proof_confidence":"verified","recommendation":"Positive ROI case: continue shadow mode and tighten attribution before expanding scope."},"evidence_register":["Annual training, delay-analysis, and workload-study exposure: $55,000,000","Primary proof metric: simulator throughput, conflict-resolution quality, workload score","Benchmark surface: run /evals before changing prompts, models, or Catalyst-Q scoring.",".rain replay: every accepted triage recommendation must carry a replay id and model/solver lineage.","Browser Run: capture screenshots for approval workflows, portal evidence, and human handoff moments.","Realtime datasets: stream optimization events through cloudflare-pipelines for finance-grade attribution."],"telemetry_contract":["event_type=roi_recommendation for every candidate recommendation","event_type=roi_acceptance for every human-approved action","event_type=roi_outcome for realized cost, service, safety, or reliability results","event_type=roi_exception for stale data, guardrail blocks, and rejected recommendations","Join keys: tenant, agent_id, workflow, rain_event_id, baseline_id, rollout_cohort, model_route, solver_version"],"acceptance_gates":["Baseline replay: Advance only if hard constraints match or beat baseline and projected value clears the buyer's pilot threshold.","Shadow mode: Advance only if operators accept the workflow and shadow deltas remain stable across normal disruptions.","Controlled rollout: Convert only the savings finance can attribute after discounting for adoption, noise, and operating cost.","Verified contract: Expand only when verified savings support the next scope and risk controls are current."],"rollout_motion":["Start with historical baseline replay and a signed baseline-cost definition.","Run shadow mode until the operator experience and deltas survive normal disruptions.","Move one bounded cohort into controlled rollout with holdout or pre/post attribution.","Renew or expand only on verified net savings after platform, pilot, and adoption discounts."],"operator_controls":["Human approval remains required for production-impacting actions.","Simulation/offline decision support first","Controller remains authority","No real-time clearance issuance","Require safety case, explainability, and certification evidence before live use"],"pricing_motion":["Free proof should show one constrained scenario, not unlimited optimization.","Paid pilot should include baseline replay, shadow mode, and the first finance packet.","Production should use hybrid pricing: base platform fee, usage meters, and optional verified outcome upside."]},"links":{"page":"https://catalyst-q.strategic-innovations.ai/roi/atc-decision-support","api":"https://catalyst-q.strategic-innovations.ai/v1/roi/atc-decision-support/packet"}}