Allometry V2

1. MVP Objective CORE

Prove one thing:

A deployment can be requested, underwritten, priced, and executed without human coordination.

The product must feel like:

• A control system — not a project management tool
• A live execution engine — not a static dashboard
• A replacement for operations teams — not a supplement

2. Core Flow (End-to-End) PIPELINE

3. Core Entities DATA MODEL

Deployment

type Deployment = {
  id: string
  accountName: string
  status: "draft" | "underwritten" | "approved" | "executing" | "completed"
  locations: Location[]
  price: number
  margin: number
}

Location

type Location = {
  id: string
  address: string
  complexity: "low" | "medium" | "high"
  units: number
  score?: Score
}

Score

type Score = {
  ev: number
  margin: number
  payback: number
  risks: string[]
}

Execution Unit (abstracted humanoid)

type ExecutionUnit = {
  id: string
  type: "unit"   // never "human"
  skills: string[]
  status: "available" | "assigned" | "busy"
  location: string
  reliability: number
}

Task

type Task = {
  id: string
  label: string
  status: "pending" | "assigned" | "in_progress" | "completed"
  assignedUnitId?: string
}

4. Routes (App Structure)

5. Screen 1 — Deployment Request INPUT

Route

/deployments/new

Purpose

Convert intent into a structured deployment request.

Layout

Components

6. Screen 2 — Underwriting + Pricing SCORING

Route

/deployments/[id]/underwrite

Layout

Components

Logic

// Margin calculation
margin = (price - cost) / price

if (margin < 0.25) → warning
if (payback > 18)   → warning

CTA

<Button>Approve & Generate Execution Plan</Button>

7. Screen 3 — Orchestration CORE

Route

/deployments/[id]/execute

Purpose

Show system replacing human coordination.

Layout

Components

Install Unit A
  → Site Prep
  → Install
  → Test

Location A
  Task 1 → Assigned → Complete
  Task 2 → Pending

Assignment Logic

function assignUnits(tasks, units) {
  return tasks.map(task => ({
    ...task,
    assignedUnitId: findBestUnit(task, units)
  }))
}

CTA

<Button>Simulate Execution</Button>
<Button>Execute Deployment</Button>

8. Screen 4 — Simulation LIVE

Route

/deployments/[id]/simulate

Purpose

Make system feel alive.

Components

Day 1 → Site Prep
Day 3 → Install
Day 5 → Complete

Simulation Engine (MVP — Fake)

function simulateExecution(tasks) {
  tasks.forEach((task, i) => {
    setTimeout(() => updateStatus(task.id, "assigned"), i * 1000)
    setTimeout(() => updateStatus(task.id, "in_progress"), i * 2000)
    setTimeout(() => updateStatus(task.id, "completed"), i * 3000)
  })
}

9. Screen 5 — Control Tower SYSTEM

Route

/control

Purpose

Show system-level power.

Layout

Components

Units Active: 42
Deployments Running: 8
Issues: 2

10. Humanoid Supply (MVP Implementation) ABSTRACTION

Internal Dataset

const executionUnits: ExecutionUnit[] = [
  {
    id: "U1",
    type: "unit",
    skills: ["install"],
    status: "available",
    location: "MTL",
    reliability: 0.92
  }
]

Assignment Logic

function findBestUnit(task, units) {
  return units.find(u =>
    u.skills.includes(task.label) &&
    u.status === "available"
  )
}

UI Rule

11. Pricing Engine (MVP) ECONOMICS

function calculatePrice(cost, risk, sla) {
  return cost * 1.3 + risk * 1000 + sla * 500
}

12. State Machine

13. Success Criteria

User must feel:

14. Demo Flow CRITICAL

1. Create deployment — enter account, locations, SLA
2. Run underwriting — automated scoring per location
3. Adjust pricing — slider with real-time margin recalc
4. Approve — generate execution plan
5. Generate execution — task tree + unit assignment
6. Simulate execution — watch it come alive

15. What Not to Build SCOPE

16. Final Positioning

This MVP communicates: