Strategy

What a COO Actually Sees: Building One Unified AI Business Operating System

I built two tools to run my business — one to plan, one to make content. The moment they worked, they forced me to admit what I was actually building: not two apps, but one operating system for the whole business.

By Logan Skees · June 20, 2026 · 6 min read

I built two tools to run my business — separately, on purpose, with my own hands and my own money, not to demo, not to pitch, but to run things. One plans. One makes and ships content. The moment they started working, they handed me a much bigger question than "what should I build next." They forced me to admit what I was actually building. This is that thinking, in the open.

TaskMaster is how I plan. Vision down to the quarter, quarter down to the week, week down to what I actually touch today. Media Machine is how I make and ship content — create it, review it, schedule it, push it live across channels. One decides what matters. The other manufactures one of the outputs that matters most. Both are in production. Both have earned their keep — on my own business first, because I don't ship anything to a client I haven't bled on myself, and then as the engine under our professional services.

Two tools, one business

So I hit a fork. Fuse them and find out how deep the rabbit hole really goes — how much of the business this thing can actually swallow — or get clever, carve out a lane where they stay politely separate, and watch the "AI OS" idea thin out into a mirage: a story I tell myself instead of a thing I build.

For a while I thought that was the decision. It isn't. I'd named the wrong fork.

TaskMaster and Media Machine were never the two things to reconcile. They're two views of one thing. And it isn't only those two — the website work, the outbound, the GTM engineering, the CRM automation are all views of the same thing. The thing is the business.

What a great COO actually does

Ask the question that settles it: what does a great COO do? A real integrator — given the people, the budget, and the authority to actually operate — what is the job? Strategy execution. Holding the entire business as one coherent object in your head and driving the plan through every function until it shows up in reality. A COO doesn't run a planning department and a content department like they're strangers who share a parking lot. He knows they're the same business expressed through different functions, and the whole value of the role is keeping them coherent.

That's the thread under all of it.

The functions are cross-mappable

The second you see it, the either/or dies. People will tell you the strategic plan lives over here, and the brand stuff — tone of voice, narrative frameworks, the offer, the content — lives over there. Different disciplines, different software, different worlds. That's the short-sighted read, and I've caught myself making it.

They're cross-mappable, link by link. My planning horizon ties to my offer. My offer carries my brand DNA — the voice, the narrative, the proof. That sets the tone for the content. The content runs on specific traffic channels. The channels feed a lead-gen system. The whole thing runs on SOPs and an IT stack. And every link throws off KPIs that feed straight back into next quarter's plan.

It's a loop, not a list of departments: strategy to offer to brand to content to channel to pipeline to numbers, and back to strategy. That loop is the business, and the loop is the COO's purview — the whole circuit, not the slices.

The real mirage: a loop that never closes

Now look at what I'd actually built. Half that loop lived in one database. The other half lived in another. Nothing joined them. Strategy, content, and the numbers each sat in their own system with no path between them, so the loop never closed. I'd built two execution systems and never put a shared model of the business underneath them.

That's the mirage. Not the ambition — the ambition is sound. The mirage is the open gap where the loop is supposed to close and doesn't. So "fusing the databases" was never a tax I pay to glue two apps together. It's the one thing that lets the circuit complete. I had a short-term view of it — I saw a migration headache where I should have seen the shared model the whole company runs on.

The mirage was never the ambition. It's the open gap where the loop is supposed to close — and doesn't.

The business is the tenant

Here's what actually changed my mind. Once you accept that the real object is the business — the whole loop — the business becomes the unit. Mine is one instance of it. Every client we serve is another instance of the same shape: their strategy, their brand, their offers, their channels, their numbers. Same model, different rows — I'm tenant number one, each client is the next.

So the other question I kept chewing on — is this a tool for the solo operator running his own shop, or a system for running many? — was a false choice too. One model, instantiated once for me and N times for everyone we work for. I dogfood as the first customer and run the rest as their fractional COO from the same chair. Solo and agency were never two products. They're one product, counted differently.

What it looks like built: engines in Focus Mode

So what does it look like? Like the architecture I already have, finally used the way it was designed.

TaskMaster has the right bones. There's an Orchestration layer — the functional areas, the org view, where you stand back and read a department's health. And there's Focus Mode, where you don't track work, you do it, one thing at a time, surfaced at the right moment. Daily planning isn't a checkbox in Focus Mode; it's a ritual you perform there.

Content marketing is no different. Making a piece, reviewing it, scheduling it — none of that is a mystery. They're discrete tasks with a predictable rhythm, which means they can surface at the right time or get pulled up on demand, exactly like planning. Media Machine's create flow, its review-and-swipe, its calendar — those don't get rebuilt. They get mounted as engines. You sit down in Focus Mode and the content engine hands you the next post to make, the next card to approve, the next slot to fill — and you do it, right there, in the same place you plan your day. Orchestration shows the same function from above: the knowledge base, the content calendar, the metrics that prove it's working.

A virtual employee for content marketing

That's not a feature. That's a hire. In plain terms it's a virtual employee for content marketing — a teammate you switch on inside TaskMaster, who shows up in your workforce, does the discrete work in Focus Mode, and reports up through Orchestration like any other function. Tied to the same strategy, the same metrics, and the same cadence as everything else you run.

The honest 20%

I won't pretend it's free. The good part — the engines, the interface, the feeling of doing content work right next to your quarterly plan — is the achievable majority, and the architecture is already shaped for it. The hard part is underneath, and I'll be exact about where: when that content engine runs inside someone's workspace, every read and write it makes — the brand it pulls, the knowledge it draws on, the calendar it touches — has to resolve back to which business this is. That's the shared model again. That's the work worth doing, because it's the only version where the loop closes.

The bet

So I can stop pretending the choice is fuse-or-mirage. The honest move is to build it incrementally — wire the engines to a single map between a workspace and a business, prove the strategy-to-content loop end to end, and let its real depth show by building into it instead of debating it from the doorway. If that loop closes — content generated from the strategy, measured against it, feeding back into it — then the AI OS was never a separate thing to chase. It's this, wired together.

That's the bet. The business is the product. Everything else is a view into it.