AI Operations Disclosure
This page documents two customer-facing AI exhibits, the data they may touch, and the controls designed around them. Runtime operation is shown only when current evidence supports it; architecture and policy declarations are identified separately.
Current operating evidence
Workshop badges require explicit runtime evidence. Historical counters, illustrative content, and static declarations do not by themselves mark an exhibit active.
forger-social-01
This exhibit documents a Facebook publishing and reply architecture and historical activity. Current publishing is confirmed by a live runtime source; reply operation is not asserted while its own source is absent or degraded.
forger-support-veloura
The public Veloura assistant currently answers supported product, shipping, sizing, return, and exchange questions. It cannot mutate orders, take payments, issue refunds, or perform account actions. Workshop conversation telemetry is connected and each figure carries its source status.
Data and controls
The architecture includes audit trails and stop controls. Their source state is shown where available; static declarations are not presented as proof that every current action is logged or every downstream routing path is operating.
Governance posture
The design emphasizes transparency, traceability, human oversight, and proportional risk management. This engineering disclosure is not a legal compliance certification. Last reviewed 17 September 2026.