ClearJanitor — Application Validation

Commercial Cleaning Robot Management

Overview

ClearJanitor is a schedule-driven, multi-tenant application that manages commercial cleaning robot fleets. Compared to WES, it exercises simpler execution contracts but introduces complex organizational relationships — leasing companies, contractors, and end users may all interact with the same robots through different access levels.

ClearJanitor validates that the platform handles recurring schedules, multi- tenant ownership hierarchies, and simpler edge contracts through the same architecture that supports WES’s complex warehouse operations. As a first- party application built by Syrius Robotics, ClearJanitor uses the same APIs available to third-party developers — no privileged access.

Ownership vs Operation

ClearJanitor supports two business models:

Model A: Leasing Company → Contractor → End User
         (ClearJanitor user = Contractor who operates the robots)

Model B: End User buys/rents directly
         (ClearJanitor user = End User)

ClearJanitor serves whoever operates the robots, regardless of ownership. The leasing company may have read-only visibility into asset utilization across all their contractors, but does not control day-to-day operations.

This ownership chain maps directly to the platform’s identity infrastructure:

Role	Platform Service	Access
Leasing Company	Org Service (management account)	Read-only visibility across contractors
Contractor	Org Service (member account)	Full operational control of assigned robots
End User	DotID (authenticated user)	Depends on Model A or B
Cross-account visibility	IAM Identity Center	Leasing company sees assets across contractors

Platform Services Consumed

Service	Plane	Usage
DeviceAdmin	Data	Enroll cleaning robots, track lease status, monitor alerts and OTA status
ThingIO	Data	Cleaning coverage heatmaps, battery analytics, robot status dashboards
OTAForge	Data	Policy-driven map and firmware distribution to cleaning fleet
DotID / StarGate	Identity	Authenticate operators, distinguish contractor vs end-user roles
Org Service	Identity	Model leasing company → contractor → end-user hierarchy
IAM Identity Center	Identity	Leasing company gets read-only visibility across contractors
Equator	Spatial	Building floor plans, zone definitions
Marie	Spatial	Coverage queries — “which floors haven’t been cleaned today?”
Planner	Intelligence	Generate cleaning routes and sequences
Scheduler	Intelligence	Recurring cleaning jobs — nightly, weekly, conditional triggers
Execution Manager	Intelligence	Monitor robots during cleaning, handle stuck/error states
Policy Service	Governance	Battery thresholds, time-of-day restrictions, zone rules
AI Policy Agent	Governance	Operators define cleaning policies conversationally

Planes Exercised

Plane	How ClearJanitor Uses It
Data Plane	Device enrollment (DeviceAdmin), coverage dashboards (ThingIO), map/firmware distribution (OTAForge)
Identity Plane	Multi-tenant access control across leasing companies, contractors, and end users
Spatial Plane	Building floor plans, coverage tracking, zone-based queries
Intelligence Plane	Route planning → recurring scheduling → execution monitoring
Governance Plane	Operator-defined cleaning rules, time restrictions, battery policies
Ecosystem Plane	Published to Marketplace, registers cleaning Domain Policy Schema on install

Validation Scenarios

1. Multi-Tenant Org Hierarchy

A leasing company owns 200 cleaning robots deployed across 5 contractors:

LeaseCo (management account)
├── CleanPro Inc. (member account) — 60 robots
├── SparkleTeam (member account) — 45 robots
├── NightShift Cleaning (member account) — 35 robots
├── FreshFloors (member account) — 30 robots
└── BrightSpace (member account) — 30 robots

• LeaseCo can view asset utilization, battery health, and lease status across all contractors via IAM Identity Center

• Each contractor manages their own robots independently

• Contractors cannot see each other’s data

What this validates:

• Org Service’s three-tier hierarchy (management account → member accounts → resources) models real-world asset ownership chains

• IAM Identity Center provides cross-account visibility without granting operational control

• The platform’s multi-tenancy is granular enough for complex B2B relationships

2. Recurring Schedule-Driven Scheduling

ClearJanitor uses recurring schedules rather than one-time order-driven assignments:

Schedule Type	Example	Frequency
Nightly	Clean all floors after business hours	Daily, 10 PM
Weekly	Deep clean restrooms	Every Sunday, 2 AM
Conditional	Clean lobby when foot traffic drops below threshold	Event-driven

What this validates:

• The Scheduler’s recurring/cron-style mode handles predictable, repeating workloads

• The same Scheduler that handles WES’s one-time assignments also handles ClearJanitor’s recurring jobs

• Conditional scheduling (future capability) is a valid use case

3. Simple Single-Path Execution Contracts

ClearJanitor contracts are simpler than WES’s multi-step picks:

Contract: Clean Floor 3, Zone A
├── Route: [waypoint sequence from Marie]
├── Coverage target: 95%
├── Failure policy: Hold and notify operator
└── Constraints: Battery > 15%, complete before 6 AM

What this validates:

• The Execution Contract model scales down gracefully — the same pattern works for simple single-path routes and complex multi-step warehouse operations

• Contract payload complexity is application-defined, not platform-imposed

4. Edge-Cloud Reconciliation (Lower Complexity)

A cleaning robot loses connectivity mid-route in a building basement:

• Projected state: 70% of floor cleaned (based on route progress and elapsed time)

• Actual state: 100% cleaned (robot moved faster than projected)

On reconnection: state updated, no replanning needed, next job may start earlier.

What this validates:

• Edge-Cloud Reconciliation works at different complexity levels — ClearJanitor’s reconciliation is simpler (single-path route) but uses the same state machine as WES

• Positive divergence (better than projected) is handled gracefully

5. Conversational Policy Creation (Cleaning Domain)

A building operator tells the AI Policy Agent:

“No cleaning between 8 AM and 7 PM on weekdays, and always prioritize restrooms over hallways”

The Agent uses ClearJanitor’s Domain Policy Schema to understand cleaning-robot, time-window, zone-priority, and creates formal policies.

What this validates:

• The same AI Policy Agent works for cleaning (ClearJanitor) and warehouse (WES) domains — Domain Policy Schemas make it truly domain-agnostic

• Different domains have different vocabulary but the same policy mechanisms

Architecture Insights Surfaced

Insight	Impact on Platform Design
Multi-tenant org hierarchy (asset owner ≠ operator)	Org Service must support ownership chains where the account operating devices is different from the account that owns them
Recurring/cron-style scheduling needed	Scheduler must support both one-time (WES) and recurring (ClearJanitor) scheduling modes
Simpler edge contracts, same model	Execution Contract pattern must scale down gracefully without overhead for simple use cases
Cross-account visibility without control	IAM Identity Center must support read-only cross-account access for asset owners who don’t operate the devices
Same execution contract model, different payload complexity	Platform defines the contract structure; applications define the payload complexity