On the Seam: Working Drafts · October 2026
The Plane Does Not Land
The transformation playbook assumes a runway. The operator has none. On modernizing while the system keeps serving customers, reskilling against a target that will not hold still, and why the exhaustion is not a sign you are doing it wrong.
I spent most of my career close to the running system, keeping it alive, carrying the pager, leading the teams that did. Except for the past nine years at the hyperscalers, where the work came wrapped in engagements: clean starts, clean stops, a scope and a kickoff and a close-out. Strategy, roadmaps, target architectures, the tidy sequence that gets an enterprise from where it is to where it should be, delivered as a thing that begins and ends. I believed in that work and I still do. But I am back close to the operators now, sitting on the customer's side of the table again, and being near them has surfaced something I am embarrassed it took leaving and returning to see.
Almost all of the advice, including plenty I used to give, is written for a company that can stop and think. That company does not exist. It is written, too, for a transformation that ends. I am less and less sure that one exists either.
The transformation playbook assumes a runway. Assess the estate, define the target state, build the strategy, reskill the team, adopt the modern practice, migrate in waves. It reads like a project with a start and a finish, and every step is sound in isolation. What the playbook never accounts for is that the person responsible for executing it is also responsible for keeping the current thing alive, today, while they do it.
You do not get to take the system down to modernize it. The system is serving customers right now. Internal ones who will file a ticket within the hour, external ones who will leave if the thing is slow or wrong. You are being asked to re-architect the aircraft while it is in the air, full of passengers, and the passengers have been promised a smooth flight and have no idea anything is being rebuilt. The plane does not land so you can work on it. There is no hangar. There was never going to be a hangar.
I watch capable people run straight into this, and the failure is almost never a failure of intelligence or will. They know what good looks like. They can tell you what the target architecture should be, why infrastructure as code matters, where the technical debt is buried, which practices they are behind on. The knowing is not the problem. The problem is that the knowing has to be acted on in the margins of a day that is already full, by a team that is already stretched, under a compliance regime that does not grant a grace period for the fact that you were busy transforming.
Compliance is the part the demos never show. In a regulated operation you cannot move fast and break things, because the things that break are the things that get audited. Every change you make has to survive a review that assumes you are guilty until documented innocent. So the modernization that looks like a weekend of good engineering in the slide deck becomes a months-long negotiation with risk, legal, and the control framework, and all of that negotiation happens while the ordinary work keeps arriving at its ordinary pace.
And then there is the reskilling, which is the quiet cruelty underneath all of it. You are told to retrain the team for the new stack, the new tools, the new way of working. Reasonable. Except the new stack is not holding still long enough to be learned. The thing you reskill toward this quarter is deprecated or displaced by the time the team is fluent in it. The half-life of a hard-won skill is now shorter than the time it takes to acquire it, which means your people are being asked to run up a staircase that is being rebuilt while they climb. They feel it. It reads as inadequacy from the inside, and it is not inadequacy. It is the arithmetic of trying to learn a moving target while delivering against a fixed one.
Underneath every piece of this is the gap I find hardest to watch, because I have stood on both sides of it, and I know which one is harder. The strategy is real and necessary, and the strategy is also the thing that gets starved, because the tactical day eats the strategic one alive. You know you need to lift your head and plan. You cannot, because the pager is going off and the release is Friday and the auditor wants the evidence by Thursday and the new hire needs onboarding and the thing you shipped last month is misbehaving in a way only you understand. The people who tell you to be more strategic have, with some frequency, never carried your pager. I say that as someone who used to do the telling.
So what do I actually have for the person living this? Not a framework. I am suspicious of anyone who hands you a five-step answer to a problem whose entire nature is that there is no step you can take without three others competing for the same hour.
What I have is a reframe, and I offer it as relief rather than instruction.
The mistake baked into all the advice is that transformation is a project. A project has an end. You push through the hard middle and you arrive, and then you operate the new thing in peace. That shape is why the difficulty feels like failure, because by the project logic you should be done by now, and you are not done, and you can feel that you will not be done. But the shape is wrong. For an operator in this environment, transformation is not a project with an end. It is a condition. It is the permanent weather now. The companies that will be fine are not the ones who finish transforming. There is no finishing. They are the ones who stop organizing around the fantasy of finishing and start building the capacity to operate continuously while changing continuously, which is a genuinely different muscle than either operating or changing alone, and almost no one is trained for it.
That reframe does not make the day easier. The pager still goes off. But it relocates the ache. You are not behind on a race that has a finish line you keep failing to reach. You are early in learning a way of working that does not have one. The exhaustion is real, but it is not evidence that you are doing it wrong. It is evidence that you are doing something genuinely hard that the people describing it have mostly never had to do.
I do not have the tidy ending here, and I am wary of pretending I do. I spent most of my career near the running system, and the past nine years delivering engagements with clean starts and clean stops, from the side of the table where the plane always looked safely on the ground. The engagement ends and you move to the next one. The operator never gets that. It took being near the work again to remember that the ground was never there. I knew that once. The least I can do, now that I know it again, is stop pretending otherwise.
The plane does not land. Build for that, and be gentler with yourself and your team than the playbook ever told you to be.
This is a working draft. Offered, not finished.
Dr. Trey Harper writes on trust, legitimacy, and the architecture of the agentic enterprise at treyharper.com and in the LinkedIn newsletter On the Seam: Working Drafts. The views expressed here are entirely my own and do not represent the policy or position of my employer or any customer.