AI Dry-Dock Scope Completeness (ADDSC)

GMI-MAR-034Maritime IndustriesIntegration-Required

Definition

Percentage of executed dry-dock work items that were present in the AI-generated scope.

The Measurement Gap

What This Metric Measures

ADDSC measures the percentage of work items in the final executed dry-dock scope that were present in the AI-generated scope. It quantifies what the AI captured during pre-dock planning versus what the superintendent, class surveyor, or yard had to add during execution: revealing the completeness of the AI's scope generation against the reality of what the vessel actually needed.

The Decision It Supports

How complete is the AI's dry-dock scope generation, and which types of work items does it systematically miss? If the AI captures 95% of the executed scope, the superintendent's review is a quick validation. If it captures 65%, the superintendent is still doing the real scoping work and the AI adds marginal value. ADDSC directly measures whether the AI scope generator is earning its place in the planning workflow.

Why Legacy Metrics Cannot See This

Legacy dry-dock performance metrics track budget variance, schedule adherence, and work item completion rate: all measured against the final approved scope. If the AI misses 30% of the work items and the superintendent adds them during planning, the final scope is correct and the budget/schedule metrics show normal performance. The AI's contribution is invisible because the human caught the gaps. Without ADDSC, the organization believes the AI is generating complete scopes because the dry-dock went well. When in reality the superintendent's experience-based additions prevented scope failures that the AI would have caused if relied upon alone.

Formula

Primary Formula

ADDSC = AI_scope_items_executed / Total_items_executed

Companion metric: AI Scope Miss Rate by category:
Miss_rate(c) = Missed_items(c) / Total_added_items for work category c

Variable Definitions

  • AI_scope_items_executed = work items that appeared in the AI-generated scope AND were executed during the dry-dock (count)
  • Total_items_executed = all work items executed during the dry-dock, regardless of whether they were in the AI scope (count)
  • c = work category (hull, machinery, class survey items, safety equipment, habitability)

Units:

Percentage. 100% means the AI scope captured everything that was ultimately executed. Track the companion Miss_rate by work category to identify systematic gaps: does the AI consistently miss hull work, machinery items, or class survey requirements?

Edge Cases

Emergent work discovered during the dock (damage found when plating is removed, conditions visible only after disassembly) is legitimately unknowable in advance. Separate "discoverable" misses (items the AI should have predicted from available data) from "emergent" additions (items that could not have been known before the dock). Track items in the AI scope that were NOT executed as a separate "over-scoping rate." Minimum measurement is per dry-dock event since each is individually consequential.

Proposed formula, pending validation pilots.

At a Glance

Unit
Percentage
Trend Direction
Higher
Reporting Frequency
Per dry-dock event
Category
Dry-Dock & Survey Optimization
← Structural Twin Prediction Accuracy (STPA)Dry-Dock Class Survey Synchronization Index (DDCSI) →

Assess Your Organization

The AI-GPS Assessment evaluates your AI governance maturity and recommends KPIs like this one for your implementation roadmap.

Take the Free AI-GPS Assessment