The skills-based organization debate is a distraction.

We're arguing about competencies vs. skills vs. tasks while skipping the question that actually matters: what are we trying to build toward?

I get it. The SBO vision is legitimately cool. Internal talent marketplace. Skills as currency. Dynamic project staffing where I can pull together a dream team based on capabilities instead of org charts. Hand me a project and a budget and let me query the system for who can help me ship it. I want that future too.

But here's what I keep running into: we're trying to build the skills layer before the foundation exists.

Tech silver bullets crash into organizational readiness. Organizational readiness crashes into poorly defined operating structures. And operating structures crash into the fact that nobody's actually defined where the F*ck the business is going. We're throwing spaghetti at walls at scale and using AI to justify why it sticks (or doesn't).

BCG said it plainly this year: these initiatives "often aren't well conceived, don't build on each other, or don't support broader workforce planning or a company's strategic goals."

No shit.

And here's the thing: we've been here before. Remember competency frameworks? (I'm betting they make a comeback in 2026, rebranded with an AI wrapper.) Competencies failed for the same reasons skills initiatives are struggling now. They were too abstract (what does "demonstrates leadership" actually mean in context?), too disconnected from business outcomes (nobody could tie them to revenue or risk), and too static (built for a world that doesn't exist anymore). Organizations spent millions building competency models that sat in binders (or SharePoint graveyards) while actual work happened around them.

Skills frameworks are making the same mistakes. Just with better software and more vendor hype.

And now we've got the task crowd. The "decompose everything into atomic units" folks who think if we just break work down small enough, we can automate our way to efficiency. This is the most seductive trap of all (especially in the AI moment). Task thinking turns humans into components to be optimized or replaced. It's the wrong frame entirely. You end up with a perfect map of what people do today (congratulations, you've documented the status quo) with no understanding of what the organization actually needs to produce tomorrow. It's rear-view mirror strategy. And it plays right into the "AI will take your job" panic instead of asking the harder question: what new capabilities should we be building?

Sangeet Paul Choudary calls this out in Reshuffle: we need to shift "from thinking in terms of tasks to thinking in terms of systems."

Task decomposition leads to human substitution thinking. Systems thinking leads to capability building. One is defensive (how do we survive AI?). The other is generative (how do we build something new?).

Here's my take: before we map skills (or tasks, or competencies, or whatever we're calling them this quarter), we need to know where leadership is actually steering the ship. And no, "cut headcount by adding AI" is not a direction (Do we need more vague AI cowbell?). That's a reaction dressed up as strategy.

Once we know the destination, we can build closed and open-loop systems to support the journey. Skills become meaningful only when they serve a defined purpose. Without that purpose, your skills taxonomy is just an expensive spreadsheet (with really nice UI). Your task inventory is just a to-do list at scale. And your competency framework is still collecting dust (now in the cloud instead of a binder).

So what do I actually mean by "systems"?

A system is a group of interacting elements that follow rules and form a coherent whole. (Boring definition, I know.) What matters is the question it forces you to ask: what is the system your company, your marketing team, your product org, your L&D function actually trying to build? How does it interact with other systems? How does it drive output against business goals? How do we orchestrate the processes, conversations, and outputs (actionable data that we could automate) in productive ways?

Without that clarity (your roadmap, your guiding principles, your organizational operating system), you're just optimizing pieces of a puzzle you haven't seen the picture for.

The irony? We finally have the tools to do this right. Pattern recognition at enterprise scale is available for $20/month (Hi [insert GPT model(s) here]). We can stop making expensive uninformed bets. We can quit guessing and start architecting.

But the tools need direction. Human work now is big-picture thinking (building the forest from the trees, not cataloging individual leaves, ... hehe that rhymed). That's what we're uniquely good at. That's what the machines can't do yet (and won't for a while, despite what the hype cycle tells you).

The fear of AI changing work is real. The exhaustion precipice is also real. Everyone's tired of the pivot-of-the-month. It's time to quit panic experimenting and start architecting.

2026 should be the year of systems thinking, solutions as a service, and figuring out direction before we chase the competency/skills/task taxonomy rabbit hole (again).

If your org is stuck in the weeds debate without a clear operating model underneath it, I'd love to hear what is happening. Drop a comment or DM me. This is exactly what makes us get out of bed at 8Levers.