# Tribal Dungeons of Global Shipping: AI Agents at Global Scale — Dmitry Buykin, Maersk

## Executive summary

This report details the engineering challenges of operationalizing complex, global business processes using AI agents. The core challenge is translating 'tribal knowledge'—which exists as unstructured Standard Operating Procedures (SOPs) or screenshots—into executable agent workflows. The speaker emphasizes that the system is not the agent loop itself, but rather the surrounding 'refining loop.' This loop involves building robust mechanisms for validation, error handling, and systematic improvement, which requires capturing a vast SOP corpus (estimated 20:1 ratio to runtime).

## Key takeaways

- The Long Tail Problem: While happy paths are often automated, the most expensive part of global operations is the 'long tail'—the exceptions and variations that require expert orchestration across multiple incomplete systems.
- Agent vs. Expert Knowledge: SOPs explain what a person sees and clicks (screenshots), but an agent SOP requires complex setup: preconditions, decisions, identifiers, backend calls, validation, recovery, and evidence of successful execution. Experts own the 'what,' and agents own the 'how.'
- The Refining Loop is the System: The true system is not the agent loop but the surrounding refining loop. This process requires making corrections executable changes, establishing guardrails to prevent unsafe paths, and ensuring behavior is observable.
- Scaling Improvement: Accuracy was not designed upfront; it was earned through systematic effort. The team logged over 100,000 corrections over nine months, using heat maps to prioritize the most beneficial work for experts and engineers.
- The Goal of AI Native Operation: AI native operation is not just about agents in a workflow; it's an adaptive architecture that learns from successful scenarios, aggregates them into composite tools, and allows repeatable rollouts across hundreds of countries.

## Technical details

- Agent Workflow Requirements: A functional agent SOP must include: preconditions, decisions, identifiers, backend calls, validation, recovery, and evidence of successful execution.
- System Architecture Components: The architecture includes four parts: SOP memory (the corpus), execution runtime, and the theme feedback capture mechanism. The SOP corpus is critical as it holds process memory modified for every country's conditions.
- Safety Mechanisms: Guardrails are essential to make 'dumb mistakes impossible.' This includes preventive measures like classifier evaluation and right gates, ensuring unsafe paths are eliminated on critical paths.
- Process Improvement Methodology: The methodology involves five steps: Make work representable; make execution bounded; make behavior observable for every agent; make correction cheap; and finally, make improvement compound.

## Practical implications

- Implementing a robust 'refining loop' is more critical than the agent model itself.
- Focusing on making corrections executable changes, rather than just opinions, drives production reliability.
- The architecture must manage variations across different countries (the SOP corpus) to achieve global scale.

## Topics

AI Agents, Process Automation, System Reliability, Global Operations, Software Engineering, Maersk Global Shipping Operations, SOP Corpus (Standard Operating Procedures)

Source: https://www.youtube.com/watch?v=dQ-_i1tZiws