Decisions and risks
| Decision | Reason | Limit / alternative |
|---|---|---|
| Fit before price | Missing approvals are a material failure even at $0 | Lowest-price fit is a prototype rule; preferences are untested |
| Paid-editor / free-viewer split | Avoid charging viewers or conflating roles | Hypothetical capacities; no actual vendor terms |
| Integer cents and bounded integers | Traceable arithmetic and explicit invalid counts | No fractional seats, taxes, foreign currencies or proration |
| Annual actual charge plus equivalent monthly | Avoid treating a yearly commitment as monthly billing | No cash-flow recommendation or real checkout |
| Fixed Studio active baseline | Consistent downgrade explanation across fictional scenarios | Scenarios change a proposed renewal team, not real customers |
| Fixed October2 clock and October31 renewal | Reproducible month-boundary outcomes | Simulation never advances active plan automatically |
| Immutable history and separate schedule | Preserve confirmed assumptions while reviewing replacements | 100-quote cap; no edit/import/export or append-only server |
| Raw saved-content comparison | Reject stale or cross-tab commitments, even with unchanged revision | Browser storage is not a transaction database; race after check remains a local-demo limitation |
| One-step assumptions Undo | Recover a deliberate draft mistake without rewriting commitments | Ends after refresh or a commitment action; cancellation/reset cannot be undone |
| Preserve unsupported saved records | Avoid silently erasing corruption or unfamiliar assumptions | Explicit reset destroys saved sample data after review |
Local storage may be unavailable, cleared or edited by the user. A warning explains memory-only work; compatible refresh is guaranteed only with working persistence. No auth, server enforcement, payments, refunds, real integrations, price optimization or willingness-to-pay evidence exists. The product boundary is visible in the UI.
Material recommendations are provisional. Software verification is separate from proposed human evaluation. Current evidence and release status: Validation.