Standing Brief

A system that only speaks when something clears its bar for interrupting you is right to be stingy about it. That restraint has a cost, and something has to pay it.

September 15, 2026 6 min read

A team wires up an assistant to interrupt them when something material changes: a decision gets contradicted, a pattern starts building, a deadline slips. Weeks go by with only the occasional message, and everyone reads the quiet stretches as reassurance. But an interrupt-only system has to be stingy on purpose, or people tune it out entirely, and stinginess means a lot of real, true, worth-knowing information never clears the bar to become a message at all. Quiet doesn't mean nothing happened. It can just as easily mean something happened that wasn't yet worth spending an interruption on.

An assistant that's earning its interruptions is, by design, leaving things unsaid

Brief's own Proactive Brief is explicit about this trade, and it's worth taking that design seriously rather than treating it as a flaw to work around: it rations itself to a couple of interruptions a week, collapses related items into one note, and stays silent by default, because an assistant that speaks constantly gets ignored regardless of how often it's right. That restraint is exactly why it's worth listening to when it does speak. It also means the bar for "worth interrupting someone about" is deliberately kept high, and everything under that bar, real signal that just isn't sharp or urgent enough yet, doesn't get surfaced through that channel. Not because the mechanism can't see it. Because surfacing it would spend a scarce budget on something that didn't earn it.

That's a different problem than an alert being too dumb to notice a pattern. A genuinely building customer theme, three small complaints edging toward one real trend, is exactly the kind of thing Proactive Brief is built to eventually catch and act on once it's confident enough to be worth someone's attention. Detection was never the limiting factor here. The limiting factor is the stretch of time before something clears that bar, plus everything that's true and worth knowing yet stays under it permanently, simply because plenty of real information is never urgent enough to earn an interruption.

The fix isn't a louder interrupt. It's a channel that owes nothing to urgency

Lowering Proactive Brief's own bar to catch more of that sub-threshold information would defeat the exact design that makes it worth listening to: interrupt more often on smaller things and the channel gets tuned out, taking the genuinely urgent messages down with it. What closes the gap instead is a second channel that was never competing for that budget in the first place: a standing brief, a full read of the current state delivered on a fixed schedule whether or not anything in it would have justified an interruption on its own. Nothing needs to be wrong, urgent, or even finished building for it to show up. That's the entire point of putting it on a schedule instead of a threshold.

Brief runs this as a separate agent rather than asking one mechanism to do both jobs. The Weekly Brief Agent, which an earlier post on this blog also looks at from a different angle, synthesizes decisions, research, and work movement into one read delivered on a fixed weekly or daily cadence, to whichever channel a team actually uses. A late interrupt and a standing brief look similar on the surface, both arrive with a summary, but they're answering different questions: an interrupt answers "did something just cross the bar," and a standing brief answers "what does the whole picture look like right now," checked on purpose instead of inferred from the absence of a message.

The honest limit: a standing brief is not a substitute for urgency

This isn't a claim that a standing brief makes Proactive Brief's interruptions less necessary, or that a weekly read is just as good as catching a real contradiction the day it happens. A decision about to be silently reversed needs same-day attention, and a report arriving on Friday is too slow for that regardless of how complete it is. The two mechanisms are solving different problems and neither one covers for the other: an interrupt is built to spend its rationed attention on the sharpest, most urgent things it finds, fast; a standing brief is built to show the full state on a schedule, including everything that was true but never urgent enough to spend an interruption on. A team with only the first misses the slow, quiet, real stuff sitting just under the bar. A team with only the second finds out about urgent things too late. Neither gap gets closed by tuning the other mechanism harder.

Think about the last stretch your team went quiet on a tracker or a decision log. Was that quiet actually confirmed, or was it just the kind of silence nobody had checked the meaning of yet?

Frequently asked questions

What is a standing brief? A read of the current state of decisions, research, and work movement, delivered on a fixed weekly or daily schedule regardless of whether anything in it was significant enough to justify an interruption. It shows up whether the period behind it was eventful or completely uneventful.

How is this different from Proactive Brief's interruptions? Proactive Brief is deliberately rationed: it only interrupts someone when a change clears a real bar for significance, because an assistant that speaks too often gets tuned out along with anything important it says. A standing brief carries no such bar. It goes out on a fixed cadence either way, which means it can surface real information that Proactive Brief hasn't yet judged worth spending an interruption on.

Why not just lower Proactive Brief's bar so it catches more? Because the bar is what makes the channel worth listening to in the first place. Loosen it and the channel starts firing on things that were never worth an interruption, people stop paying attention to it, and the messages that actually mattered get lost in the same noise. A separate, scheduled channel can carry the sub-threshold information without ever touching how tightly Proactive Brief is rationed.

Does a standing brief replace Proactive Brief? No, they're built to answer different questions and neither one substitutes for the other. A contradiction that needs same-day action has to interrupt someone; a weekly cadence is too slow for that. Something real but not yet urgent enough to earn an interruption has no natural home in a system built to ration itself, which is exactly the gap a fixed-schedule report is built to fill instead.