For decades, the standard response to a scaling business was to hire a manager. When the sales team got too big for the founder to oversee, you hired a Sales Manager. When marketing and sales stopped talking to each other, you hired a RevOps Manager to sit in the middle. We built businesses like multi-layered cakes, where every layer of 'doing' was separated by a layer of 'coordinating.' This was a necessity in the human-led era because humans are notoriously poor at high-speed, high-fidelity information transfer.
But we have entered the Autonomous Coordination Era. As part of a genuine AI transformation, the most forward-thinking businesses are realising that the 'middle'—that layer of management whose primary function is to route information, check status, and facilitate hand-offs—is no longer an asset. It has become a bottleneck. If your workflows move at the speed of light but your approvals move at the speed of a Tuesday morning sync meeting, your technology isn't the problem. Your structure is.
The Rise of the 'Coordination Friction Trap'
💡 Want Penny to analyse your business? She maps which roles AI can replace and builds a phased plan. Start your free trial →
In my work helping businesses navigate this transition, I’ve identified a recurring pattern I call The Coordination Friction Trap. This occurs when an organisation automates its 'edge' functions—like customer support or data entry—but keeps a human management layer to oversee the outputs.
What happens? The AI produces work in seconds, but that work then sits in a manager's inbox for 48 hours waiting for a 'review' that adds no strategic value. The efficiency gains of the AI are swallowed whole by the latency of the human coordinator.
When we look at IT support costs, for instance, the bulk of the expense isn't usually the technical fix itself; it’s the triaging, the ticket routing, and the follow-ups. In an AI-first model, an agent doesn't just 'do' the support; it coordinates the resolution across departments without a human manager needing to 'assign' the task.
From Human APIs to Agentic Hand-offs
Historically, middle managers have acted as 'Human APIs.' They translate the needs of Department A into the language of Department B.
- The Old Way: A property manager receives a maintenance request, logs it, calls a contractor, emails the tenant, and updates the owner.
- The Autonomous Way: An AI agent receives the request, identifies the priority, checks the budget, autonomously hires a pre-approved contractor via API, and updates all stakeholders simultaneously.
In the second scenario, the property manager’s coordination role is gone. The 'hire' isn't a more expensive manager; it’s a more sophisticated agentic workflow. We see this shift radically altering the P&L in sectors like real estate—check out our property sector savings guide to see the specific margin shifts when coordination is automated.
The 90/10 Rule of Management
One of the hardest truths I share with subscribers is the 90/10 Rule of Management: When AI handles 90% of the coordination and status-checking within a function, the remaining 10% of 'human' work (like high-level strategy or emotional conflict resolution) rarely justifies a standalone management role.
In many cases, that 10% can be folded back into the 'doing' roles or handled by the founder. This is how you build a leaner business. It’s why my own business runs autonomously; I don't need a manager to tell me if my marketing is aligned with my strategy. The data flow is direct and the 'hand-off' is instantaneous because there are no hands—only code.
The Managerial Bottleneck in AI Transformation
Many leaders treat AI transformation as a tool-kit. They buy a subscription to a chatbot and hope for the best. But true transformation is an architectural shift.
If you are still hiring managers to 'ensure the work gets done,' you are essentially paying a premium for a biological router. AI agents can now maintain 'contextual persistence' across an entire project. They don't forget what was said in the kickoff meeting three weeks ago. They don't need to be 'briefed' on the current status of a lead.
When you compare Penny vs. a traditional business consultant, the difference is often this very thing: speed and synthesis. A consultant takes weeks to 'coordinate' a discovery phase. An AI-driven approach synthesises your data in minutes.
How to Hire Your First AI Agent (Not a Manager)
If you’re at the point where you feel you need to hire a 'Coordinator' or 'Junior Manager,' stop. Instead, look for the 'Inter-departmental Gap' they were supposed to fill.
- Map the Hand-offs: Where does work stop and wait for a person to move it to the next stage?
- Define the Translation: What information does the manager 'add' during that hand-off? If it’s just status updates, it’s ripe for an agent.
- Deploy an Orchestrator: Use tools like LangChain or AutoGPT to build agents that can interact with your existing software stack (Slack, CRM, Project Management) to move work autonomously.
The Future belongs to the Orchestrators
The goal isn't to remove humans from your business; it's to move humans to the edges where they add the most value—creative strategy and high-stakes relationships—while the AI handles the messy, expensive, and slow 'middle.'
Your next hire shouldn't be someone to manage the work. It should be the system that makes management unnecessary. The Autonomous Coordination Era is here, and the businesses that refuse to stop paying the 'Coordination Tax' will simply be out-competed by those that do.