Flattening Your Org Chart Is Not a Transformation Strategy

Coinbase made headlines this week by laying off 14% of its workforce and restructuring its leadership around "player-coaches" — leaders who are also strong individual contributors, overseeing larger teams and moving faster with AI at the center of how work gets done.

Armstrong framed it plainly: the company is not just reducing headcount. It is fundamentally changing how it operates.

He is right that something fundamental is changing. But the Coinbase story — and the wave of similar announcements from Block, Snap, Meta, and others — is being read as a template when it is actually a case study. And case studies only transfer when the context does too.

The Part of the Story That Is Not Being Told

Every restructuring announcement describes the after state. The flatter org chart. The new team structure. The AI-native pods.

What is rarely described is the work that has to happen for people to actually operate that way.

Flattening a leadership structure is a decision that takes an afternoon. Building the capability, the culture, and the operating model that makes it functional takes months. Most organizations — particularly large, complex ones with multiple business units, established workflows, and thousands of employees — are attempting the second part without a clear architecture for how to get there.

That is where transformations fail. Not in the announcement. In the absorption.

What Workforce Transformation Actually Requires

Successful AI-driven workforce transformation requires four things in place — and in the right sequence.

*Change management built around people, not communications.* Most organizations treat change management as a communications problem. Real change management starts with understanding where the resistance lives and building a deliberate architecture to address it before the rollout begins. This is harder than it sounds — pinpointing resistance requires genuine buy-in from line managers, who often have the clearest view of where friction will emerge but the least incentive to surface it early. The difference between 40% adoption and 95% adoption is almost always in the upstream work, not the training.

*Workforce transformation tied to actual role redesign — including re-skilling.* AI does not just change what tools people use. It changes what their jobs are. Organizations that skip this step end up with employees using AI for the wrong things, or not using it at all, because the workflow was never redesigned around the technology. The tool gets blamed. The real problem was the process — and the path forward is still being written differently in every firm. There is no template yet. This is a virgin path, which is why structured re-skilling and training must be built into the transformation from the start, not bolted on afterward.

*Stakeholder management that runs in sequence, not in parallel.* Champions get activated first. Managers get equipped second. Broad communication comes third. Reversing that sequence is one of the most reliable ways to generate resistance where there was none.

*Process and adoption architecture that embeds the tool in the flow of work.* When a tool is embedded in how work actually gets done — not sitting alongside it as an optional resource — adoption compounds naturally. The process redesign work has to precede the training, not follow it.

What Coinbase Actually Demonstrates

Armstrong's restructuring is not a model for how to transform a workforce. It is evidence of what is possible when a workforce has already been through the transformation. The engineers who adopted new tools within a week were operating in a culture that had been building toward this for years.

For years, Armstrong has been all in on AI. After securing GitHub Copilot and Cursor licenses for every engineer, he went rogue, asking engineers to get onboarded with the tools by the end of the week rather than the quarters some in the company had said it would take. Those who did not meet the deadline faced consequences.

That is one way to do it. It works when you have a founder who controls the culture, a technical workforce that adapts quickly, and a board that will absorb the short-term disruption.

But even for Coinbase, we do not yet know how this will work out. It is entirely possible that the talent forced through this transition leaves at the first opportunity. It is also not clear whether this is genuine transformation or a marketing narrative — similar to what other tech leaders have done: overhire, then restructure and call it AI-driven efficiency. A complex product like Coinbase, dealing with crypto custody and compliance, relying heavily on AI without clear precedent, is a riskier proposition than the headlines suggest.

Most Organizations Are Not Coinbase

Most organizations have thousands of people across multiple business units, each with its own culture, its own workflows, and its own relationship with technology. You cannot mandate adoption by Friday and fire the holdouts. You have to build the architecture that makes adoption possible — and then move people through it.

That architecture has four components that every organization navigating AI transformation needs to get right. If done in earnest, this approach works — but it takes time, and keeping a lid on what is happening internally while the transformation unfolds is a real challenge. Leaks create premature narratives. Narratives create resistance. The organizations that manage this well treat internal communication timing as seriously as they treat the transformation itself.

1. Change management. Not communications. Not a town hall. A structured approach to how resistance gets identified, addressed, and converted into momentum.

2. Workforce transformation. Which roles change, which workflows get redesigned, and who gets upskilled for what. This is not an HR question. It is an operating model question. And it must include deliberate re-skilling — because the roles that remain will require fundamentally different capabilities than the roles that existed before.

3. Stakeholder management. The executives who need to sponsor this, the managers who need to model it, and the employees who need to trust it. Getting all three aligned — at the right time, in the right sequence — is what separates adoption from compliance.

4. Process and adoption architecture. The new flows, the new orchestration, the new ways of working that make the technology stick. Not because people were told to use it. Because it is now embedded in how the work actually gets done.

The Organizations That Will Win

The leaders who navigate this successfully are not the ones who move fastest to restructure. They are the ones who build the architecture that allows their people to absorb the change, redesign their processes, and move with genuine speed — not because they were told to, but because the system makes it possible.

Armstrong is right that AI is bringing a profound shift in how companies operate. What he is describing at Coinbase is the output of that shift. What he is not describing is the work it takes to get there.

That is what an organizational operating system does. And it is what distinguishes a transformation from a restructuring.

*GeneralArc designs and implements the change architecture that allows complex organizations to absorb AI-driven transformation — from readiness assessment through adoption.*