Practice

The Day the Matter Resurfaces

A matter never becomes sensitive on the day it is handled. It becomes so months later, when someone asks how the decision was made.

Imagine an opinion delivered on a Tuesday in March. Nothing spectacular: a structuring question on a mid-sized transaction, handled seriously by a competent team, validated by a partner, sent to the client at day’s end. The matter closes. The next day, everyone has moved on, because that is how a firm breathes: you decide, you deliver, you move to the next. No one thinks about that opinion anymore.

And it is precisely at that moment, when everyone has turned the page, that the only life cycle that will truly matter begins, the one no one watches. The March opinion did not die on being delivered; it entered a long latency, during which it sleeps in a directory, apparently closed, in reality only suspended. Its real story has not yet begun, and nothing, at the moment it is closed, lets one guess when or how it will begin.

A delivered matter is not a closed matter. It is a matter awaiting a question not yet asked.

The hidden life of a delivered matter

One spontaneously reasons as if a matter lasted as long as its handling: you open it, you work it, you deliver it, it is over. This representation is false, and it is the source of the risk. A delivered matter does not cease to exist; it merely changes state. It passes from an active object, which one handles, to a dormant object, which one no longer watches but which remains fully liable to be recalled. Its real lifespan is not that of its handling, but that of its consequences, which can extend over years.

This hidden life is poorly accounted for because it is silent. During the months the matter sleeps, nothing happens that draws attention; no alert recalls that it might one day resurface. The firm lives at the rhythm of its active matters, and the immense mass of its dormant matters occupies no place in its daily attention. Yet it is in this silent mass that the risk resides, precisely because one no longer thinks of it. One protects what one watches; one neglects what one has ceased to see.

One must therefore operate a reversal of perspective on what a matter is. Not a piece of work that ends at delivery, but a commitment that runs as long as its effects, and whose most exposed part is not the one handled actively, but the one filed away and forgotten. The most dangerous moment of a matter is not the one where it is produced, under the team’s vigilant eye; it is the one, much later, where no one watches it anymore and where it can be recalled without warning.

Fourteen months later

Fourteen months later, the context has changed. The client’s company is the subject of an acquisition, or an audit, or a dispute with a third party. Suddenly, that quiet March matter finds itself in the spotlight. An acquirer launches its due diligence and combs through past decisions. A regulator questions the structuring adopted. An insurer, before covering, demands elements. And the question that recurs, from the mouths of all these interlocutors, is invariably the same, posed almost word for word: how was this decision made, on the basis of what elements, by whom, and with what verifications?

For any legal or compliance professional, this question is nothing extraordinary. It is even routine: it is the daily life of a matter that gains importance. What is new is not the question, it is the difficulty of answering it when the opinion was prepared with a tool that kept nothing of its own making. For when that March opinion was produced with a copilot’s help, the chain that led to it dissolved at the very moment of its production. One knows, vaguely, that a tool was used. One knows roughly when. One ignores all the rest.

One must measure that this is not a problem of the opinion’s quality. The opinion can be excellent, perfectly founded, perfectly defensible on the merits. The problem is elsewhere, in the gap of time: what was a shared obviousness at the moment of producing it, in March, has become, fourteen months later, an impossible reconstruction. The context that made the decision clear has faded from memories; the people have changed matters, sometimes firms; and there remains only a correct document, but orphaned of everything that would explain how it was reached.

An opinion can be right in March and indefensible in May of the following year, without having changed a comma.

The risk that does not announce itself

What makes this risk particular is that it does not distribute equally among matters, and above all that it does not announce itself. Ninety-nine opinions in a hundred will never be reopened, and for those, the absence of a record will cost nothing. But one does not know, at the moment of producing, which will be the hundredth. It is the very definition of a latent risk: it is invisible until the day it is not, and on that day, it is too late to constitute the proof one should have kept.

This temporality is the heart of the problem, and it distinguishes it from anything one can anticipate. A risk you see coming, you handle as it approaches. A latent risk gives no warning: the matter that resurfaces sends no signal in March, at the moment one could still act. When the signal arrives, fourteen months later, the window to protect oneself has long since closed. One cannot decide, after the fact, to have been prudent before.

This temporal asymmetry has a consequence few organizations draw: since one does not know in advance which matter will resurface, the only protection consists in treating every matter as if it could resurface. Not from excess of caution, but from simple lucidity about the nature of the risk. One cannot sort sensitive matters in advance, because sensitivity is not a property of the matter at the moment it is handled; it is a property the future assigns to it, retrospectively, according to events that have not yet occurred.

You do not know in advance the matter that will resurface. The only protection is to have held every matter as if it would.

Protecting yourself before knowing you will need it

Yet a tool that works in isolated interactions cannot meet this requirement, for a simple reason: it never held the continuity while it worked, and one does not reconstruct after the fact a continuity that never existed. At the moment the question arrives, it is already too late; either the memory of the making was constituted in real time, in the course of execution, or it will never exist. Protection against the resurfacing matter therefore cannot be added on the day it resurfaces: it must have been put in place months earlier, blind, without knowing it would serve.

This is what distinguishes a governed execution layer from a generation tool. In MAX, the memory of the making is not an optional module one would activate for matters deemed sensitive; it is the operating condition of the layer itself, present on all matters, including the ninety-nine that will never resurface. It is the price, and the only possible form, of a protection against a risk that does not announce itself: it must cover everyone, since one ignores whom it will protect.

One must grasp what this inversion changes in the firm’s position the day the question falls. Without this memory, it finds itself summoned to prove, after the fact, the rigor of a process of which it kept no imprint: an almost impossible exercise, always costly, deeply anxiety-inducing for the teams who must undertake it under urgency. With it, the situation is reversed: what allows answering existed before the question was even posed. The firm does not plead from an absence, it answers from something held at the right moment, that is, before knowing it would matter.

This is the whole meaning of a layer designed for duration rather than for the instant. It does not serve to make opinions better on the day they are delivered; it serves to make them defensible on the unpredictable day someone reopens them. At the moment the March opinion is produced, nothing signals that it will one day become the heart of a high-stakes discussion. It is precisely for that reason that it must be treated, from the first day, as if it could.

The layer does not make the opinion better on the day it is delivered. It makes it defensible on the unpredictable day it resurfaces.