What You’ll Learn
In this tutorial, James walks through the complete AI Operating System he built for TrainingSites.io — including 3 boards that now manage his entire teaching business. You’ll see how he uses Claude to write custom WordPress plugins, track AI agent activity, and run a Kanban board that his AI employees manage on his behalf.
This isn’t about prompting AI better. It’s about making a fundamental shift: treating Claude as an employee you give structured instructions to — not a chatbot you ask questions of.
The Big Shift: From Plugins to AI Employees
If you’ve been running a WordPress site, you know the drill. Find a plugin. Try it. It doesn’t do exactly what you need. Try six more. Pay for three of them. None of them fit perfectly.
James hit the same wall — and then flipped the model entirely.
“I’ve been writing WordPress plugins — or at least extending WordPress as a content management software — to do what I wanted it to do, as opposed to purchasing plugins.”
Instead of buying a plugin and adapting to it, he describes what he needs to Claude and Claude builds it. The Skills Library on TrainingSites.io (252 skills)? Built by Claude. The Agent Bundles plugin? Claude. The tutorial system, the calendar — all built through conversation, in 4–5 iterations.
💡 In Plain English: Think of Claude as a developer on your team. You describe what you need, it builds a first version, you give feedback, it revises. Four or five rounds later, you have exactly what you wanted — and you own it.
Board 1: The Agent Logger
The first board James shows is the Agent Logger — a WordPress dashboard that tracks every AI agent running in his business.
For each agent run it captures:
- Which skill or agent ran
- Whether it succeeded, partially completed, or errored
- How much time it saved
- Approximate cost saved
This gives James a real-time picture of what’s working. Not a hunch — actual data. When something breaks, he sees it immediately. When an agent is saving 4 hours a week, he sees that too.
✓ Check Your Work: Do you know which AI tasks are running in your business right now? If the answer is “not really,” the Agent Logger concept is your starting point.
Board 2: Mission Control
The second board is Mission Control — also built directly inside WordPress, connected to Claude and all running agents.
It shows James, at a glance:
- Which scheduled tasks are running and which aren’t
- What’s blocked and needs attention
- Recent activity — what got done this week
- All active MCP connectors (the bridges between Claude and external tools)
“I actually have a mission control board now that is directly within WordPress. It is hooked to Claude and all of the agents that are running.”
The key here isn’t the technology — it’s the operating posture. James isn’t logging into five different dashboards to see what’s happening. Everything surfaces in one place, and anything that needs his decision is flagged for him.
💡 In Plain English: Mission Control is like a cockpit. You’re not flying the plane manually. You’re watching the instruments and stepping in only when something needs a pilot.
Board 3: The Kanban Board
The third board is where things get really interesting. James has a Kanban board (FluentBoards) inside WordPress, and Claude manages it on his behalf.
The columns:
- Ideas — new work coming in
- Waiting on James — agent did its part, needs human input
- In Progress — agents actively working
- Blocked — needs technical resolution
- Done This Week — completed work
At the start of every session, Claude reads the board, flags what needs James’s attention, and picks up where it left off. James doesn’t need to remember what was in flight — the board holds that context.
“I can work with Claude here and actually have him run my business from the Kanban board.”
✓ Check Your Work: Is there a single place where you and your AI can both see what’s in progress, what’s waiting, and what’s done? If not, you’re losing continuity between sessions.
Why This Matters for Educators in 2026
The old model was: you do the work, AI helps occasionally. The new model is: AI does the work, you direct it.
That shift sounds simple. But it requires you to set up the infrastructure that makes it possible — the context files (instructions, ICP, brand, systems), the memory files, the boards. Without those, every Claude session starts from zero.
James puts it plainly:
“It’s about giving your agents instructions on what tasks you want completed. This has nothing to do with prompts whatsoever.”
For educators, coaches, and consultants, this matters because your time is spent in the room — facilitating, coaching, creating transformation. The business operations that surround that? Content repurposing, email campaigns, community engagement, offer marketing — those can run with agents while you focus on what only you can do.
⚠️ Advanced: Building this full system takes time to set up correctly. If you’re just starting, focus on one board first. The Agent Logger is the easiest entry point — it gives you visibility without requiring you to change how you work.
What’s Coming Next
James mentions this video is the teaser. The follow-up videos will break down each component in detail:
- How Mission Control is built and connected
- How the Agent Logger captures data from Claude sessions
- How the Kanban board workflow runs day-to-day
- The org chart for AI departments and how James interfaces with each
If you want to follow along as he builds, join the community free at trainingsites.io/join.
Key Takeaways
- Build, don’t buy. Claude can write custom WordPress plugins that do exactly what you need — no compromises.
- Visibility first. Before you can manage AI employees, you need to see what they’re doing. The Agent Logger solves this.
- One board for everything. Mission Control + Kanban in one place means no context-switching and no lost threads.
- Structured instructions, not prompts. The shift from “asking Claude things” to “giving Claude a role and a task list” is what makes the difference.
- Your job changes. When agents handle operations, you operate as CEO — setting direction, reviewing decisions, stepping in when needed.
