The hidden tax

Modern companies are built around functional silos: product, engineering, marketing, sales, finance.

But every important initiative is cross-functional. That contradiction creates an enormous hidden tax inside organisations.

AI may fundamentally change that.

Companies scaled through specialisation. That made sense when human expertise was scarce and expensive.

The problem is that specialisation also created coordination overhead:

• translation layers
• fragmented context
• communication latency
• knowledge attrition

Every time work crosses a function, information degrades. Intent gets compressed. Context gets lost. Trade-offs get reinterpreted.

The final outcome is often worse than the sum of the parts.

Large organisations behave like distributed systems with high communication latency between nodes. And as companies scale, that coordination tax compounds.

When coordination becomes the bottleneck

For decades, this trade-off was unavoidable. Specialisation increased efficiency even if it slowed coordination.

But AI changes the economics that made those structures necessary in the first place.

Most people think about AI vertically:

• engineers code faster
• marketers produce content faster
• sales teams generate pipeline faster

But the more important shift may be horizontal. AI reduces the cost of moving across functions.

One person can increasingly:

• think
• write
• prototype
• analyse
• design
• code
• execute

inside the same feedback loop.

The distance between intention and execution collapses.

And that dramatically reduces organisational packet loss.

From idea to execution

Over the past few months, I've repeatedly found myself bypassing traditional functional dependencies while launching a new software product inside the business.

Instead of writing specs for someone else, waiting for prioritisation, translating intent across teams and coordinating multiple specialists, I could increasingly move directly from idea → execution myself.

Not perfectly.

But fast enough that the coordination cost became more expensive than the execution itself.

That's the shift I think many organisations are underestimating: AI doesn't just make specialists faster. It reduces the friction between domains.

The AI-native operator

If AI continues reducing the cost of traversing functions, companies may need:

• fewer coordination layers
• fewer handoffs
• fewer rigid functional boundaries

The highest-leverage operators won't necessarily be the deepest specialists. They'll be the people who can integrate across domains with speed, judgment, and AI-native execution.

For decades, org charts were designed around the limits of human specialisation.

AI may force us to redesign companies around a very different constraint: the speed at which context can move from intention to execution.