AI Can Draft the Org Chart. It Can’t Build the Infrastructure.
- Jules Siegel-Hawley
- Aug 5
- 3 min read
Updated: Aug 13
Recently my focus has shifted from the shape of organizations to what actually holds them together, especially in fast-moving, product-driven environments.
Infrastructure (i.e., the scaffolding that keeps things moving when priorities shift, decisions stall, or everything feels urgent all at once) is easy to overlook when everything’s flowing. But when it’s not, even the most elegant of operating models can start to break down -- and start-ups certainly feel this faster than most.
The Outputs Are Getting Better, But The System Still Feels Off
AI is rapidly improving our ability to generate the formal artifacts of organizational life: team structures, role descriptions, process documentation, even proposed decision flows. The surface of the organization has never looked more polished. And more disconnected from how the work actually happens. I’ve seen this firsthand in teams where AI tools are humming, but cross-functional ownership and flow haven’t caught up.
I keep hearing the same frustrations from leaders: decisions that don’t stick, priorities that shift without warning, teams that move fast but not always in the same direction.
The technology has gotten smarter, but the outcomes are lagging. And this disconnect points to a deeper truth: the issue isn’t just what the org looks like, but how the org functions. And more often than not, that’s a question of infrastructure -- the kind that doesn't show up in dashboards.
The Work Under the Work
Organizational infrastructure is what makes strategy executable. It’s how decisions actually get made, how work flows across functions, how ownership is distributed, and how people know what’s expected of them, especially when there isn't an obvious answer.
It includes the systems most teams don’t talk about until they break: decision rights, prioritization mechanisms, escalation paths, rituals that reinforce alignment, and the informal norms that shape how people interact when things get messy.
AI can help document these systems. It can even suggest improvements. But it doesn’t know how your organization actually runs.
It can’t see that a handful of leaders are holding everything together through sheer force of will.
(Sound familiar?)
It doesn’t recognize the unspoken dependencies or the legacy decisions that no one has questioned in years. And it doesn’t pick up on the hesitation in a room when something important is left unsaid.
AI is obviously useful, but it is, for the time being at least, working with a limited view. It can capture the output, but not necessarily the behavior behind it.
AI Will Mirror Your Reality, Not Necessarily Improve It
There’s a growing belief that AI can help fix the way we work, and that if we just feed it the right information, it will give us clarity in return. But AI doesn’t challenge your assumptions; it reflects them.
If your organization is already running on workarounds, bottlenecks, or unspoken hierarchies (i.e., organizational debt), those patterns won’t disappear. Instead, they’ll be mirrored in your tools -- more efficiently, more elegantly, but just as misaligned.
Clean process documentation doesn’t equal operational clarity, a beautifully rendered RACI chart doesn’t mean people feel safe to make decisions, and overall faster output doesn’t solve for misalignment.
When we automate without interrogating the underlying system, we risk scaling the confusion rather than solving it.
What Still Belongs to The Humans
This is the tension I keep coming back to. AI can handle so much of the visible, repeatable work. But the invisible system—the one that governs how your organization actually operates—still requires human judgment, discipline, and care.
At Andes, we work with organizations to surface what’s often overlooked: the assumptions behind the workflows, the habits that shape how decisions get made, the handoffs no one ever formally designed. We ask the kinds of questions that don’t show up in dashboards, but that determine whether your strategy has a chance of landing.
What are you optimizing for right now?
Do your systems reflect that, or are they still serving a previous phase?
Where are you relying on tools to do what leadership never clarified?
What parts of your infrastructure are silently carrying risk?
Where to Begin
If your team is moving fast but feels uncoordinated, it may be time to pause. Not to slow down progress, but to check what kind of system you’re building.
Is it designed to scale what matters most? Or is it simply documenting what’s already misaligned?
Even in an AI world, the fundamentals haven’t changed. Organizations still run on trust, clarity, and shared understanding. Tools can certainly reflect and support those things, but they can’t create them.
That’s still human work. And in moments of inflection -- a product launch, a hiring sprint, a big shift in direction -- it’s the infrastructure that determines whether the momentum holds.
Infrastructure doesn’t build itself. It has to be shaped, maintained, and led by people who are willing to take ownership of how the organization really works.
