Table of contents
Open Table of contents
The thing AI actually changed
When people talk about what AI did to companies, they reach for the obvious story: it made execution faster. Code gets written in hours. Landing pages, ad variants, campaigns, docs — all of it ships at a speed that would’ve looked like science fiction three years ago.
That’s true. But it’s the boring half of the story. The interesting half is the second-order effect almost nobody is pricing in: when execution gets this cheap, the bottleneck slams straight into leadership.
Building was never the expensive part of being right. Deciding what to build was. And AI didn’t touch that — it just stripped away everything that used to hide it.
The clock got faster on the wrong people
Here’s the uncomfortable math. If building a product used to take nine months and now takes six weeks, you didn’t just get faster — you compressed the time leadership has to make the strategic call before the resources are committed.
In the old world, a slow execution cycle was an accidental safety net. A bad strategic decision took a year to fully play out, which gave you a year to notice, course-correct, and quietly absorb the mistake. The sheer cost of building bought you time to think.
That net is gone. When a team can build and launch the wrong thing in six weeks, a wrong strategic bet is now a six-week-old fact in the market — fully shipped, fully marketed, fully wrong. The penalty for a bad decision arrives faster, and you have to make more decisions to keep the faster machine fed. More at-bats, shorter clock, higher stakes per swing.
And it isn’t just engineering that compressed. The entire go-to-market motion did. AI now does the deep research that used to gate sales — qualifying leads, profiling accounts, drafting tailored outreach, surfacing the right talking point for the right buyer in seconds instead of days. Customer support compressed the same way: AI triages, researches, and resolves the bulk of tickets before a human is ever paged. So it’s not one function that got faster — it’s build, market, sell, and support, all at once. Every part of the company that used to act as a natural buffer between a strategic decision and its consequences got thinner. When research-and-respond collapses across the whole value chain, there’s almost nothing left between a leader’s call and the market’s verdict on it.
Speed is now a commodity — direction isn’t
This is the part leaders need to sit with. If everyone can execute fast, then execution speed stops being a competitive advantage. It becomes table stakes — the cost of staying in the game, not a way to win it.
Your competitor has the same models, the same agents, the same compressed cycles. They can build the feature you’re building just as fast. So what’s left to compete on?
The only durable advantage left is pointing the fast machine in the right direction. Two companies with identical execution speed will diverge entirely based on the quality of the bets their leadership makes. One ships three brilliant things and two duds a quarter; the other ships five duds. Same velocity. Wildly different outcomes. The difference is entirely upstream, in the judgment.
When everyone can build fast, building fast is no longer the edge. The edge is knowing what to build — and being willing to bet on it before you’re certain.
I’ve made a version of this argument about engineers: when implementation gets cheap, the scarce skills become specification, taste, and verification. The exact same shift happens one level up. For leadership, the scarce skill is strategic judgment under a compressed clock — and it’s now the highest-leverage thing in the company.
What this demands from leaders
A faster machine doesn’t forgive a slow or fuzzy hand on the wheel. A few things change concretely:
- Decision velocity becomes a core competency. You can no longer afford a six-week deliberation to authorize a six-week build. The decision-making cadence has to match the execution cadence, or your fast teams sit idle waiting on a slow table.
- Conviction under uncertainty matters more. You’ll have less time to gather perfect information before committing. The leaders who win are the ones who can make a high-quality call on incomplete data and adjust — not the ones who wait for certainty that never arrives.
- The cost of indecision inverts. It used to be safe to wait. Now, while you deliberate, a competitor has already shipped, learned, and iterated twice. Standing still is the new risky move.
- Strategy becomes a continuous act, not an annual offsite. Quarterly planning was built for slow execution. When you can change direction in weeks, strategy has to be a living, frequent discipline — sensing, deciding, redeploying.
Where the real moat is now
None of this means decisions should get reckless. The opposite. Faster execution raises the value of being right and the cost of being wrong simultaneously, which means the quality of thinking behind each decision matters more than it ever has — you just have to produce that quality faster.
This is why I keep arguing that the organisational brain is the real moat. When you have to decide quickly and well, the differentiator is the depth of context you can bring to the call — the accumulated decisions, the dead ends already explored, the institutional memory that lets a leader make a sharp judgment in an afternoon that would take a competitor a month. Fast execution makes good context more valuable, because it shortens the distance between a decision and its consequences.
AI didn’t make leadership less important. It did something more demanding: it took away the slow execution that used to cushion bad strategy, and handed leaders a faster, sharper, less forgiving instrument. The companies that win the next decade won’t be the ones that execute fastest — almost everyone will execute fast. They’ll be the ones whose leadership can think clearly, decide quickly, and bet well on the compressed clock.
Execution got cheap. That just made the thing leaders do the most expensive — and the most valuable — work in the building. This is the bet behind my whole AI Engineering practice: the leverage moved up the stack, all the way to the top.