Automation applied to a broken process doesn’t fix it. It just runs it faster.
When researchers set out to actually quantify bureaucracy rather than complain about it, one number kept surfacing: non-managerial employees spend around six and a half hours of every week simply complying with internal process. Reports nobody reads, approvals that route through four people, reviews that exist because of the organization rather than because of the work. Scaled up, the estimate is that excess bureaucracy costs the U.S. economy more than $3 trillion a year in lost output, and close to $9 trillion across the OECD. Those hours never appear as a line item. They appear as slowness, and they get blamed on almost everything except their actual cause.
That’s the layer AI is now being installed onto. Between February and March 2026, Grant Thornton surveyed 950 business leaders across ten industries. Seventy-eight percent lacked strong confidence their organization could pass an independent AI governance audit within ninety days — meaning most couldn’t show how AI decisions get made or who owns the outcome. Forty-eight percent of boards had approved major AI investments without setting governance expectations at all. And only 6% of executives named change leadership and workforce enablement as a top skill needed to thrive in an AI-driven environment.
That last number is the one to sit with. Organizations are funding the technology and underfunding almost everything that determines whether it works: process redesign, decision rights, training, the structural questions nobody wants to own. The same survey found organizations with fully integrated AI were nearly four times likelier to report revenue growth than those still piloting — 58% against 15% — and traced the gap not to better models but to clearer accountability. Which means the six and a half hours don’t disappear when the software arrives. They just run faster.
Gary Hamel and Michele Zanini spend all of Humanocracy on that gap — between the technology a company buys and the operating structure it gets installed into — and the six-and-a-half-hour figure is theirs. Their argument is that the constraint on most large organizations isn’t strategy, talent, or tooling. It’s bureaucracy: layers of approval, rigid role definitions, and control systems designed for a nineteenth-century problem that almost every company still runs on.
What makes the argument useful rather than merely damning is their insistence that it’s measurable. Their Bureaucratic Mass Index scores an organization on things leaders rarely quantify — how many layers sit between the front line and the CEO, how much time goes to internal compliance rather than external work, how quickly a bold idea gets killed, how much personal advancement depends on political skill. That’s what produces a number a board can act on. “We have too much red tape” is not.
Two ideas worth carrying into the boardroom: Treat the operating model as the first line item, not the last. If the constraint is structural, adding software to it buys speed inside a process that shouldn’t run at all — and the faster it runs, the more expensive the underlying design becomes to unwind. And measure the drag before debating the fix. Bureaucracy survives because its costs are diffuse and invisible while its benefits are legible and defended. Nothing changes the conversation faster than a specific number attached to a specific process, which is why Hamel and Zanini built an index rather than an argument.
Read Humanocracy →
Best for: Executives who suspect their organization’s problem is structural rather than technological. Reading commitment: Around seven hours, dense and research-heavy, with substantial case material.
Seeing exactly where the organization stalls, instead of arguing about where it’s slow, is what Celonis is built to deliver. It’s a process intelligence platform that connects to the systems a company already runs on and reconstructs how work actually moves through them: every approval, every rework loop, every handoff that sat idle for three days. Rather than the process as documented, it shows the process as executed, which is almost never the same thing. Named a Leader in the 2026 Gartner Magic Quadrant for Digital Twin of an Organization Platforms, its position is that visibility is only the starting point — the platform is designed to feed that live picture of operations into AI and automation as context, so the automation is aimed at a process someone has actually looked at.
It’s a heavy lift, and honest about who it’s for. Reviewers consistently flag a steep learning curve, real dependence on data quality from source systems, and customization that requires specialists rather than business users. It also won’t tell you which processes deserve to exist — it will show you, in detail, that an approval chain has six steps, but the decision to remove three of them is still a leadership one that software cannot make for you. For organizations about to spend heavily on AI, though, it answers the question that should come first: what does the process actually look like before we make it faster?
There’s also an opportunity attached to this issue: an exclusive feature placement in Business Insider, arranged through TEI’s media relationships. Business Insider carries particular weight with employees and prospective hires — the people who experience an organization’s approval layers daily and can tell immediately whether a company operates the way it says it does, which is exactly the gap this issue is about. TEI’s team handles the interview, drafting, and placement end to end.
If that’s of interest, reply to this email or grab time on our Calendly and we’ll cover the details.
Take with you: Before your next AI investment gets approved, does anyone in the room know how many steps the process it’s meant to accelerate actually has?
The Executive Insight — 169 Madison Ave STE 11570, New York, NY, 10016
