18. Staged Execution Plan

Timing adjusts to funding, institutional cooperation, and regulatory review. Each phase is evidence for deciding whether to fund the next.

Phase 1 · Months 0–3: Daily requests and initial companion

Build conversational mobile, web, and phone access; two or three simulated or public-service connections; accessibility profiles; regulated payment approval links; failure and handoff; user-configured conversation and contact order; and small co-design sessions.

Entry criterion: users can express the goal and verify outcomes in simulated requests.

Phase 2 · Months 4–6: Secure execution space

Build managed isolated browsers, encrypted workspaces, one-time authority, approval across channels, passkey and attestation pilots, long-running booking checks, outcome proof, and deletion receipts.

Entry criterion: complete bounded booking and change requests without a personal PC, duplicate action, or residual sensitive data.

Phase 3 · Months 7–10: AI kiosk field trial

Build a lightweight kiosk prototype, QR/NFC session, disability-specific personalization, pre-payment confirmation, staff call, automatic deletion, and trials at two to five sites.

Entry criterion: improve completion for accessibility users while reducing store burden.

Phase 4 · Months 11–14: Companion, community Nodes, and rewards

Operate N0–N4 trust levels, fixed companion and telepresence trials, community and specialist Nodes, quality measurement, check-in and false-alarm measurement, KRW fees plus performance-token trials, helper matching, and disputes.

Entry criterion: demonstrate per-request revenue and cost without privacy violations.

Phase 5 · Months 15–20: Institution links and mobile-robot research

Run booking and visit-preparation trials, review MyHealthWay, FHIR, and KR Core, apply field-level health consent, assess institutional privacy and security, and research bounded telepresence and mobility under safety, insurance, ergonomics, and physical-stop review.

Entry criterion: agree legal and technical responsibility and validate safety in a real workflow.

Phase 6 · Month 21 and beyond: Safe robotics and service stability

Expand only after external Node-security audit, accessibility governance, token-contract and reward audit, long-term token-economy validation against real revenue and Node cost, and safety, regulatory, insurance, and field evidence for any physical assistance.

Not prioritized initially

IROA will not initially build its own foundation model, diagnostic medical AI, full hospital-record replacement, its own stablecoin, unrestricted mobile-app control, face-recognition personalization, continuous household recording, unattended care replacing people, or physical-contact robotics before safety validation.