Most agency leaders I work with are convinced the chaos comes from talent gaps, culture issues, or individual behavior. After 200+ engagements, I can tell you: it almost never does. The real cause is how the work itself is structured and coordinated — and until you see that, you’ll keep solving the wrong problem.
And it's about to get worse.
The Hidden Tax of Instability
Here’s a scene I’ve seen dozens of times. A team istruggles with misalignment, and rework keeps happening, scope drifts, clients are frustrated. The natural response? More communication. More Slack channels. More standups. More check-ins. More status meetings.
And the Chaos grows.
Not surprising when you understand what’s actually happening. Agencies are what I call NOCOs, Naturally-Occurring Chaotic Organizations. They run multi-client, multi-project, multi-allocated operations where any given worker might serve 5+ different clients with 8+ different managers who all think their project is the priority. The default state of these organizations isn’t order — it’s chaos. Not because of bad people, but because of structural complexity.
Recent research supports what I've been saying and teaching for more than a decade now: Lauring and Jonasson’s work on hybrid and distributed teams shows that when work becomes asynchronous and people shift between modalities (ex: remote to in-person, Slack to Zoom to email) informal alignment doesn’t just weaken, it changes form entirely. Coordination has to be rebuilt through deliberate structure, not left to lightweight ambient interactions that existed in times of lower chaos (Lauring & Jonasson, 2025).
The Natural Reaction That Makes Things Worse
Here’s the thing: adding more communication when things feel chaotic is a completely natural reaction. Something’s not working, so you increase the signal. You add a standup, a new channel, add a weekly sync. It’s what any reasonable person would do.
But natural doesn’t mean effective. As I wrote in MIT Sloan Management Review, it is very easay to overmanage complicated collaborations, and the additional meetings and structure that result put additional stressors on productivity rather than relieving them (Skeels, 2025). A single 30-minute meeting might seem like a small tax on efficiency. But when meetings are multiplied across overlapping projects, they result in hours of context-switching and lost productive time. Time card hours look full, projects appear to be staffed, everybody says they’re busy — but in reality, people are treading water.
When a team of eight sits in a standup where a manager talks first, assigns priorities, and asks for status — that’s not coordination. That’s surveillance. The workers already had a plan for their day. That meeting just destroyed it. Every interruption to a knowledge worker in flow costs at least 15 minutes of recovery time. A manager who walks the floor interrupting 40 people has just destroyed 10 hours of productive time in 20 minutes of “morale boosting.”
Recent research on communication in hybrid environments confirms this pattern: effective coordination requires structured practices and intentional rhythms, not simply more messages across more channels (Zhu, 2025). The agencies that get this right don’t communicate more. They communicate deliberately, at the right times, in the right structure, with the right people in the room.
And Then AI Enters the Picture
AI fits this pattern perfectly. It’s the latest version of the same natural reaction: things feel slow and chaotic, so you reach for a tool that promises speed.
And AI does deliver speed..but only at the action layer. It generates drafts faster, produces more variations, automates routing and formatting. But action has rarely been what holds organizations back. The bottleneck is the judgment layer: interpreting context, aligning priorities, weighing tradeoffs, determining what’s good enough to proceed. And judgment cost doesn’t scale linearly with organizational size. As it scales, it become less efficient: every additional person, team, or stakeholder doesn’t just add one more judgment call — it multiplies the coordination and alignment required across the whole system. (I’ve written about this at length in The Bottleneck AI Can’t Fix and of course, my book. )
AI creates more output, produced faster, flowing into a judgement system that was already struggling to keep up. More deliverables competing for the same review bottlenecks. More versions requiring the same overstretched managers to evaluate and align.
Organizations adopting AI often feel simultaneously faster and more stuck: They’ve accelerated the part that wasn’t the problem. The structural friction that actually governs their speed — how understanding is shared, how coordination happens, how decisions get made — remains untouched. Until that changes, AI just amplifies the chaos.
Don't miss out on new articles!
Get every new article delivered to your inbox — no spam, just ideas worth reading.
Why Culture Alone Isn’t Enough
When adding more communication doesn’t work — and adding AI tools doesn’t work — leaders often reach for the next natural lever: culture. They remind people of the mission. They reinforce values. They hope that a renewed sense of shared purpose will help the team persevere through the chaos.
Again, this isn’t a bad instinct. Culture supports psychological safety, motivation, and trust...pretty much all things that matter. But culture cannot fix design failures in coordination structures. If your people are multi-allocated across six clients, interrupted twelve times before lunch, and working in a management structure where ten people can interrupt any one person on any given day — no amount of purpose-reminding or team-building will fix the operational dysfunction underneath.
Leaders lean on culture because structural dysfunction feels like a “people problem.” When things aren’t working, it’s natural to assume that better attitudes, more engagement, or stronger commitment will close the gap. But the research on team cognition tells a different story: structural clarity is foundational to performance. Teams need shared mental models — a collective understanding of what the work is, who’s doing what, and how it all fits together — and those don’t emerge from good vibes. They emerge from structure.
Recent meta-analytic evidence across dozens of studies shows that shared mental models are positively correlated with both team process and performance — and critically, the way those models are built matters as much as whether they exist at all (DeChurch & Mesmer-Magnus, 2010).
Some agencies believe that if the office feels fun and loose — skateboards in the hallway, beer fridges, casual Fridays — they've built a strong culture. But that's decoration, not culture. Culture is how the organization actually behaves under pressure, and no amount of office perks will fix a coordination structure that has ten people interrupting any one person on any given day.
Structural Fixes Leaders Can Make: What Works
After years of testing this across agencies of every size, I’ve found that the highest-leverage move is protecting the daily structure of work. Not big reorganizations. Not new project management tools. The day itself.
The Check-In and Day Structure Model works by inverting the typical meeting cadence. Instead of mandating quiet times (which never stick), you prescribe “noisy times” — dedicated windows for meetings, check-ins, and collaborative discussion. Everything outside those windows is protected for flow. Workers go first in check-ins. Managers ask questions; they don’t direct. Nobody is “in charge” of the meeting — it exists to serve the team, not to serve management’s need for visibility.
This typically eliminates half of existing meetings. Not through policy. Through structure. When you designate the right times for coordination, the unnecessary meetings become visibly unnecessary.
Pair this with pod structures — small, cross-discipline groups organized around a focused cluster of clients — and you dramatically reduce the number of connections and management layers any one person has to navigate. The chaos doesn’t go away entirely. But it becomes manageable. And the productivity gains are real: across our client data, agencies that address these structural factors see 30–50% faster delivery and 40+ point improvements in client satisfaction.
And I suspect you realize already: better delivery makes for a better culture.
Go Deeper
I’m running two free webinars on March 10 (and 17) that dig into exactly these dynamics:
“Why Agencies Become Crazy Places to Work — And Why AI Makes It Worse” (8:00 AM PDT) covers the NOCO model and the structural sources of agency chaos — including why AI is amplifying these problems, not fixing them.
“The Skateboard Culture Myth — Why AI Won’t Fix Your Culture Problem” (3:00 PM PDT) challenges the assumption that culture is the lever for performance and explores what actually drives motivation and inclusion in structurally complex organizations.
Both are live, interactive, and free. If anything in this article made you nod — or wince — I’d love to see you there.
