Observation
Warrant: Derivation
Two curves moving at different speeds, and the gap between them is being spent.
The lag between what the technology can do and what leadership believes it can do is real, it widens every month, it is temporary, and it is the resource this firm is currently working inside.
01 · Two curves
One curve has accelerated. The other one never did.
What a frontier system can do is moving faster than almost anything in economic history. The capability available to anyone who cares to use it changes on a timescale of months, and the change is not incremental in the way most technology change is incremental. That is the first curve, and it is the one everybody is looking at.
The second curve is what leadership inside a large industrial company believes that capability to be. It is set by how quickly a conviction travels through an organisation: who has to be persuaded, what has to be seen, which committee meets when, how long a budget cycle runs, how many people have to be wrong in public before a category is admitted to exist. That curve is bound by ordinary human and bureaucratic speed, and ordinary human and bureaucratic speed has not accelerated at all.
Neither curve is a criticism of anyone. An industrial organisation is built to be slow to change its mind, and that property is load bearing: it is the same property that stops a plant reorganising itself around every passing claim. The point is only arithmetic. When one quantity accelerates and the quantity that tracks it does not, the distance between them grows, and it grows without anyone deciding that it should.
That distance is the subject of this page. Not the technology, and not the organisation. The distance.
02 · The gap is the resource, and it decays
Warrant: Conditional
A resource that is consumed by being correct about it.
The gap is not a permanent condition to design a company around. It is stock, not income. Every month it is wider, and every month some of it is spent, because belief eventually arrives whether or not anyone helps it along. A plan that assumes the distance stays this wide is a plan built on something being used up. The correct response is not to enjoy the gap. It is to convert it, while it is open, into the one thing that does not decay: a record of contact with reality that nobody can reconstruct later by reasoning harder. That conversion is the whole argument of the trajectory, not the assets.
03 · Why ground level entry does not work
A vendor list is a defence, and it is doing its job.
The obvious way into an industrial company is the level where the work happens: the site manager, the maintenance lead, the procurement officer who signs for the services already being bought. It is the shortest route to a conversation and it is the wrong one, for a reason that has nothing to do with the people in those roles.
Those roles exist structurally to defend an existing vendor list and an existing budget line. That is what they are for. A procurement function that admitted every new category on its merits would not be a procurement function. Someone at that level who is persuaded has been handed a problem rather than an opportunity: a thing they now believe in, with no line item to buy it against, and a personal cost to advocating for it that is paid whether or not it works.
So the conversation is not lost on the argument. It is lost on the org chart. A category with no existing line item can only be authorised where line items are created, and line items are not created at the level where they are spent. This is the same structural point made about advisory work at advisory work is priced as though reasoning were scarce: the shape of the buying decision, not the quality of the reasoning, decides what can be bought.
04 · What only leadership can authorise
Warrant: Foundation
The access the record needs is not the access that is easy to get.
There is a second reason to enter at leadership altitude, and it is the more important of the two, because it is about what the work is for rather than about who signs.
A ground level relationship grants access to routine, already well understood work. That work is real and it is worth doing, but it is thin in the only currency that matters here: it produces observations of situations the record already covers. The rare, cross site, information rich deployment, the one that permanently extends what is known, exists only where somebody can authorise access across sites, across contractors and across the boundaries that normally keep each observation sealed inside the job it came from. That authority does not sit at the site.
Every engagement is therefore two things at once, a source of revenue and a source of information, and the second is the one that is scarce. The standing rule for choosing between them is set out at how work is chosen.
| What is being asked for | Where it can be authorised |
|---|---|
| Routine work of a kind already well described | At the site, inside an existing line |
| A repeat of a task the record already covers | At the site, inside an existing line |
| The same task compared across several sites | Only above the site |
| A deployment nobody has instrumented before | Only above the site |
| Outcomes returned to the pooled record afterwards | Only above the site |
05 · Not closed by technical education
The anxiety already arrived, and it did not arrive as a technology story.
The instinct, once the gap is named, is to try to close it by explanation: brief the leadership of an industrial company on what the current systems do, install the technical understanding first, and sell into the understanding afterwards. That sequence is wrong, and it is wrong twice.
It is wrong first because nobody at that altitude needs to understand the machinery to be receptive, any more than they need to understand a balance sheet's audit trail to know what an unreliable one costs them. It is wrong second because the relevant anxiety is already present, fully formed, and it arrived through a channel leadership already reads. Assurance work that a company paid an outside party to perform, and that later turned out not to have been performed in the way it was represented, is not a story about technology. It is a story about whether the account you were given of your own operations was true.
That is the doubt worth attaching to, because it is the doubt that is already there. The question it raises is not which tools a supplier uses. It is a narrower and more uncomfortable one: what, structurally, would stop this supplier being wrong without anyone finding out. Everything this firm publishes about who pays, about liability, and about revisiting its own past claims is an answer to that question and not to the other one. Start at who pays, then the calibration ledger.
A trust story is answered with structure. A technology story is answered with a demonstration. Only one of those two survives contact with a sceptical room.
06 · The ask, framed at the real maturity
Early in a long build, and said in those words.
There is an obvious temptation at leadership altitude, and it is the one thing on this page that is a decision rather than an observation. The temptation is to describe the system as further along than it is: the definitive answer, already, for a whole portfolio. That framing invites exactly the scrutiny an early record cannot yet survive, and it invites it from the most sophisticated audience available.
So the ask is framed at the maturity the firm actually has. A real, working system. Early in a build measured in years. A seat in shaping what is built next, held by the counterparties who are inside it first. That is a smaller sentence than the alternative and a more durable one, and it is the only version of the sentence that stays true as the record grows underneath it.
The reasoning is not modesty. Overselling maturity does not cost a contract when it collapses. It costs the account of this firm that circulates afterwards in a small, relationship driven network of people who all speak to each other, and that loss is larger and much slower to heal than any single deal. The general form of the argument, and why this firm's domain punishes an unchecked claim faster than any other, is at stated limits.
- 01 The system is real and it works Stated in the present tense, with no claim attached about scale, coverage or outcomes already checked.
- 02 It is early in a long build Said first, by the firm, rather than discovered later by the counterparty.
- 03 The offer includes shaping what is built next The counterparties inside it early are inside the design of it, not only inside the price of it.
- 04 Nothing checkable is claimed ahead of being true Not accuracy, not recognition, not a result. Those sentences are earned in one order and no other.
The gap between the two curves closes on its own schedule and takes nobody's permission to do it. What is done inside it, while it is open, is the only part of this that is a choice.