The power they shut off
An automated shutoff can cut the power to a house on a billing flag without anyone deciding that this house, with these people in it, should go dark — and by the time a human looks, the harm from a decision no person made has already been done.
Consider a household whose electricity goes off on a weekday afternoon. There was no knock, no final call answered, no person who looked at the account and decided that this was the day. A balance crossed a threshold — a payment that posted late, or a meter reading a system misjudged, or a charge disputed weeks earlier and never resolved — and a rule did what rules do. It sent a signal down the line, a switch opened, and the house went quiet. Somewhere in the record there is a timestamp and a status code. There is no moment where a human weighed what the loss of power would mean here, in this house, with the people who happen to live in it.
What makes the case worth sitting with is not that a mistake might have occurred. Mistakes occur everywhere. It is that the disconnection is a physical act with physical stakes — light, refrigeration, a device that has to stay charged, heat or cooling in weather that can be dangerous — and it was carried out with the same ceremony a spreadsheet uses to close a row. The family discovers the loss the way you discover any sudden absence: the fan stops, the screen goes dark, the food in the refrigerator begins its slow turn. Then begins the second ordeal, which is trying to get a human being to reverse in hours what a rule accomplished in a second.
The house that went dark by rule
The disconnection executed on a flag. That is the whole of it, and it is worth being precise about what the flag did and did not contain. It contained an account number and an amount and a category — arrears, or a meter exception, or a failed autopay. It did not contain the fact that someone in the house depends on a powered medical device, or that there is an infant, or that the outside temperature that afternoon made the loss of cooling a health event rather than an inconvenience. It did not contain the earlier phone call in which the bill was disputed, because the dispute lived in one system and the shutoff logic read from another. The rule saw a number over a line. It did not see a home.
This is the ordinary shape of harm at scale. No individual chose to darken this particular house; the choice was made once, in the abstract, when someone wrote the policy that says balances past a threshold are disconnected, and thereafter it executed itself against every account that matched. The person who wrote the policy never met this family and could not have. The system that applied it could not weigh what it could not see. And so the harm arrives before any human is in a position to prevent it — which is the precise inversion of how a decision this consequential ought to run. The judgment came after the act, if it came at all, and by then the account of a disconnection was easier to produce than the electricity.
A physical act treated like a ledger entry
There is an older version of this decision, and it is worth naming so we are honest about what automation removed rather than merely romantic about the past. A person used to stand between the overdue balance and the open switch. That person was imperfect, sometimes arbitrary, and slow. But the role that person played was to hold the specific circumstance up against the general rule and, occasionally, to decline — to notice the medical device, the newborn, the heat advisory, the dispute that was plainly still open, and to stay the shutoff for a case that should not proceed. The discretion was the safeguard. Automating the disconnection did not just make the rule faster; it deleted the moment where a circumstance could interrupt it.
The result is that a dangerous act is now performed with the affect of a routine one. A ledger entry is reversible and hurts no one; you correct it and the harm was only ever notional. A disconnection is neither. It lands hardest, and this is not incidental, on the household least equipped to absorb it and least able to contest it quickly — the family without a second account to draw on, without a day free to spend on hold, without the fluency to escalate past the first line of a phone tree. The people a discretionary pause most protected are exactly the people an automated rule most exposes. The stakes and the speed run in opposite directions, and the automation widened the gap.
Shutting off a home's power is a physical act with physical consequences, and a system that performs it as casually as it closes a ledger has confused a household with a balance.
None of this is an argument that nonpayment must be tolerated, or that a utility has no legitimate interest in collecting what it is owed. It plainly does, and a service that cannot manage delinquency cannot stay solvent enough to serve anyone. The argument is narrower and harder to wave off: the interest in managing a balance does not license treating the means of managing it — a dangerous physical act against a specific home — as though it carried no more weight than the balance itself. The account may be routine. The shutoff is not.
A check before the act and an account after
What accountability requires here has a shape, and it is not the abolition of collection. It is a decision built to the stakes of what it does. Before the switch opens, the decision owes a check for the circumstances that should stay its hand — the registered medical need, the protected weather window, the dispute still formally open, the household flag that says this one gets a human before it gets a disconnection. These are not exotic categories. They are the very circumstances the discretionary role used to catch, now demoted to fields a rule declined to read. A pre-execution check that consults them is the difference between a system that enforces a policy and one that enforces it blindly.
And after — if a shutoff does proceed — the decision owes a fast, legible account and a fast remedy, held to the tempo of the harm rather than the tempo of a billing cycle. A household in the dark cannot wait out a thirty-day review; the reason it went dark should be stateable in plain terms on the day it happened, and a human with the authority to restore power should be reachable at the speed the danger demands. This is where a Decision Receipt does honest work: not as a courtesy, but as the record that makes the disconnection contestable while contesting it still matters — what flag fired, what circumstances were and were not checked, who can reverse it and how fast. The party with the most at stake, and the least room to wait, is owed standing to challenge the determination before the cold sets in, not after.
The line to hold is simple to say and easy to erode. Managing nonpayment accountably is a legitimate thing to build. Disconnecting a home the way you close a spreadsheet row is not the same thing, however identical the two may look inside the system that does them. A decision that can endanger a household owes a check that considers the household before it acts, and an account that answers to the household after — because the alternative is a dangerous act performed with a routine act's indifference, on the people least able to survive the mistake.
The scenario above is illustrative — a composite drawn to show a pattern, not an account of any real person, company, or event.
— Dispatches · Summit Cognitive
Continue from here
Turn the argument into a practice.
Get new dispatches, assess how your organization handles consequential decisions, or explore Summit Cognitive.