For decades, the logistics and supply chain sector has operated on a simple, albeit expensive, equation: volume equals headcount. If you wanted to move more freight, manage more carriers, or oversee more complex international routes, you hired more coordinators. You scaled your back office in lockstep with your bill of lading. This created what I call the Coordination Tax—a structural inefficiency where 30% or more of a logistics firm's margin is consumed simply by humans acting as the 'glue' between disparate software systems.
Today, we are witnessing a fundamental break in that equation. The AI transformation of the back office isn't just about 'automation' in the old sense of if-this-then-that rules. We are moving into the era of the Agentic Back Office, where autonomous AI agents handle the reasoning, the negotiation, and the exception handling that once required rooms full of people.
I’ve seen this pattern emerging across hundreds of the businesses I advise. The result? Global logistics operations that once required fifty people are now being run by two-person teams. These aren't just 'tech' companies; they are traditional operators who have realised that in 2026, human capital should be used for strategy, while AI handles the execution.
From Static Automation to Agentic Agency
💡 Want Penny to analyse your business? She maps which roles AI can replace and builds a phased plan. Start your free trial →
To understand why this shift is so disruptive, we have to distinguish between 'legacy automation' and 'agentic workflows'.
Legacy automation is brittle. It’s a series of hard-coded rules. If a shipment is delayed by 24 hours, the system sends an automated email. But if the shipment is delayed, the alternative port is congested, and the customer’s contract has a specific penalty clause for weekend deliveries? Legacy automation breaks. A human has to step in, check three different screens, call a broker, and make a decision.
Agentic AI is different. An agent isn't just a script; it’s a goal-oriented entity. When you give an AI agent a goal—'Minimise the impact of the Port of Felixstowe delay on our Q3 margins'—it can:
- Parse the data: Read the incoming delay notification.
- Reason through the context: Check the manufacturing schedule to see which components are critical.
- Explore alternatives: Query carrier APIs for available space on alternative routes.
- Execute the solution: Re-book the freight and update the customer via their preferred channel.
This is the end of the 'Coordination Tax'. The human doesn't do the work; the human monitors the agent doing the work.
The Architecture of a Two-Person Global Logistics Team
How does a two-person team actually manage a global flow of goods? They don't do it by working harder; they do it by becoming Agent Orchestrators. In this model, the two humans divide the business into two roles: the Architect and the Guardian.
1. The Architect (Strategy & Integration)
The Architect focuses on the 'top-down' view. Their job is to ensure the AI agents have the right 'tools' to work with. This means managing API connections between the warehouse management system (WMS), the ERP, and external carrier platforms. They are constantly looking at fleet management costs and asking: 'Is our AI agent making the same decisions a high-level procurement manager would?'
2. The Guardian (Exception & Ethics)
Even the best AI agents hit 'The 90/10 Rule'. AI can handle 90% of logistics permutations perfectly. The remaining 10% are the 'black swan' events—geopolitical shifts, sudden insolvency of a major carrier, or ethical dilemmas that require human judgement. The Guardian steps in only when the AI flags a high-uncertainty event.
By narrowing the human focus to only the most complex 10%, a single person can oversee a volume of work that would have previously overwhelmed a department of twenty.
The Death of the 'Agency Tax' in Logistics
For years, many businesses outsourced their logistics to third-party providers (3PLs) not because the shipping itself was hard, but because the coordination was too complex to handle in-house. This is the Agency Tax—paying a premium for someone else's headcount.
As AI agents become more accessible, we’re seeing a massive trend toward 'Insourcing'. Companies are realising they can run their own sophisticated logistics desk using an AI-first approach. When you look at the savings available in transport and logistics, the biggest line item isn't fuel or rubber—it's the administrative overhead of 3PLs who are still using 2015-era human processes.
The 'Elastic Back Office' Framework
If you are a business owner looking at this shift, you need a framework for adoption. I recommend the Elastic Back Office model, which involves three phases:
Phase 1: The Digitisation of Intent
Before an AI can act for you, it has to understand your 'intent'. Most logistics knowledge is trapped in the heads of senior staff. You must document your decision-making logic. 'We always choose Carrier X for perishable goods unless the delay is over 6 hours'—this is intent. Agentic AI needs this as its 'North Star'.
Phase 2: Tool-Equipping
You don't just 'install' AI. You give it access. This means moving away from closed, legacy systems to platforms with robust APIs. If your AI can’t 'see' your inventory or 'talk' to your hauliers, it's blind and mute.
Phase 3: Shadow Agency
Run your AI agents in 'Shadow Mode'. Let them suggest decisions for 30 days without executing them. Compare the agent's decision to your human team's decision. Once the agent hits 95% alignment, you 'cut the cord' and let it execute.
Why the Window is Closing
The competitive advantage of the two-person logistics team isn't just that they are cheaper. It's that they are faster.
An AI agent doesn't sleep. It doesn't wait for a Monday morning meeting to re-route a ship that was diverted on a Saturday night. In a global economy where supply chains are increasingly volatile, speed is the only real hedge against chaos. The businesses that move toward an agentic back office now are building a level of resilience that their headcount-heavy competitors simply cannot match.
I’ve worked with thousands of businesses, and I can tell you this: the transition from 'human-led' to 'agent-led' is the most significant commercial shift of the decade. It’s not about losing the 'human touch'—it's about using the human touch where it actually adds value, rather than using it as a high-priced patch for broken processes.
Are you still paying the Coordination Tax, or are you ready to build your two-person global operation?