---
title: "Why Our Flagged Items Queue Grew Every Single Week"
canonical: https://dxdev.com/everyday-ai/the-review-pile-that-can-never-hit-zero/
datePublished: 2026-08-03
---
We set up a review pile the obvious way. Anything uncertain, anything that didn't quite fit the normal pattern, anything a process couldn't confidently handle on its own, got routed there. Then we set the rule we thought any responsible person would set: the system isn't trustworthy until that pile is empty.

That rule was wrong. We only found out because someone else looked at the plan before we'd built it and talked us out of it.

## Why "clear it to zero" sounds responsible

Nobody argues with "don't let things pile up, clear them, stay on top of it." It sounds like discipline. We were already picturing where the file would live and the rule that would sit at the top of it: not trustworthy until empty.

The trouble is that a pile which has to hit zero before you can trust anything is competing with everything else you're supposed to do that day. It doesn't stay small on its own. It becomes a second inbox, and unlike a real inbox, nobody else is around to help sort it.

## The question that changed our minds

Does this pile get gated on reaching zero, or does it fade out on its own when nobody has touched an item in a while?

For a small team, the answer was this. If nobody has looked at a low-stakes flagged item within a set window, it probably wasn't important enough to need looking at. Requiring zero before you trust the system will paralyze the person running it. Letting everything sit forever untouched turns the pile into a weight that only grows. Letting items age out keeps the pile useful: a live view of what's recent and might actually need attention, with everything else falling into the archive.

That changes what the pile is for. It is a spotlight on what's unresolved and still recent. It was never meant to be a permanent ledger of everything ever unresolved. Once an item ages past the window, it stops being a task hanging over you and becomes a record. Nobody has to close it, because closing was never really the point.

## We had already built this trap once

We had built the same trap before without noticing. Every uncertain item that didn't meet a confidence bar went to the same pile, with no expiration. Left alone, that number never shrinks. It grows week over week for as long as the underlying process keeps running.

An expiry rule changes the shape of the problem. Instead of an ever-growing list you're supposed to feel guilty about, you get a rolling window that shows what's current. An item that sat untouched for the agreed window wasn't important enough to touch. It ages out un-reviewed, and that's a fine outcome. The record still exists if anyone needs to go back to it, and the person running the pile stops being its permanent bottleneck.

The same goes for anything you stockpile "just in case": old notes, flagged exceptions, things you meant to get back to. If something solidifies into a real pattern worth acting on, it graduates into your actual process. If it goes untouched long enough, let it fall away instead of sitting there demanding attention it was never going to get.

## What the assistant does and what a person decides

An AI assistant is useful for designing this kind of pile and proposing a sensible window based on how the work actually flows. It can also flag which categories are too high-stakes to ever let expire: anything touching money, a customer relationship, or a legal obligation. It shouldn't decide on its own that an item is safe to let go, and it shouldn't delete or archive anything without a person confirming first.

A person decides which piles need a must-be-empty rule, the ones where an unreviewed item is a real risk, and which are safe to let fade. That's a judgment call about what matters, and it stays with you.

Decide which of your piles must be empty because an unreviewed item is a real risk, and let the rest age out.