Here is what actually happens when a global company announces a new strategy. The CEO stands up and names the new direction. The executive team spends the next one to three months translating it into something their functions can act on. Each function's leaders take another month or two to derive their version. Directors turn those into plans, which takes two to four more. Somewhere around month eleven, the strategy has technically cascaded. Then comes activation: who owns which piece, who approves what, and, after the reorganization that usually rides along, who do I even talk to now. Call it seventeen months before the new strategy is genuinely being worked at the level where work happens.
The board does not wait seventeen months. Around month six the questions start: you announced a new direction, where are the results? But strategic results are lagging indicators. The honest answer, that the strategy has not fully reached the people doing the work yet, is not an answer a boardroom rewards. So leadership brings in a brand-name firm to diagnose why the strategy is not working. The firm recommends a new one. The cycle restarts.
I watched this from inside one of the world's most complex global companies, through four reorganizations in six years. The part nobody writes down: the strategy that got replaced was often working. It just had not had time to show up in the numbers. And every restart, some of the people who carry the institutional knowledge decide they have seen this movie before, and leave.
It is not a people problem
The frustrating part is that the knowledge to move faster already exists inside the building. The planner three levels down knows things about how the business actually runs that no executive deck will ever capture: where the real bottlenecks are, which handoffs break, which assumptions in the diagram stopped being true two years ago. Strategy that ignores that layer is strategy built on how the company is supposed to work instead of how it does.
The cascade is slow because it is a game of translation, played top-down, one level at a time, with the people who hold the ground truth brought in last, if at all.
Inheritance, not translation
Now imagine the cascade as inheritance instead. The strategy is announced once. Within days, every level of the company, from the executive team down to the individual contributor, receives an aligned first draft of its own part in it: derived from the level above, informed by what that team actually does and knows. Nobody starts from a blank page, and nobody waits four months for the level above to finish translating.
Then the important part happens: people push back. The planner flags the assumption that does not survive contact with reality. The regional lead says this sequencing breaks our market. Those corrections do not die in a meeting; they propagate, up to the levels whose plans depend on them, down to the teams they redirect, and sideways to the functions they touch. The strategy inherits from below and directs from the top, at the same time.
This is the opposite of replacing your people with AI. It leverages the humans the company already employs, all of them, as the strategy engine. The machine does the propagation, the drafting, and the consistency checking. The humans do the knowing, the judging, and the steering. Every level of management and every individual contributor, contributing to one comprehensive strategy that is alive at every altitude of the company.
Seventeen months to two
Run the same announcement through that company and the math changes kind, not degree. A global organization that used to need a year and a half to truly change direction can do it in less than two months. A company that size, turning essentially on a dime, without pretending the org chart is the company.
Calafai today runs one engagement at a time: sourced, challenged, graded strategy work for the people closest to the problem. That is the foundation, and it is real now. The cascade is where it is going. If you run a company and this picture bothers you in a good way, we should talk.