Your agents don’t need to persist. Your purpose does.
The backlog grooming, through Unl
This archived article remains available at its original address. Unl, your why agent, carries the decisions you settle and why into your AI work. See how it works.
Backlog grooming is meant to keep the list honest — a working set of things actually worth doing, not an archive nobody has the heart to close. Most groomings skim the newest cards and leave the old ones alone, because judging whether an old card still matters takes more than a skim.
A backlog that only grows is a backlog nobody’s actually curating. Unl holds the standard you ratified — a card stays only if it still serves a live objective — so grooming can check the old cards as rigorously as the new ones and name what’s actually dead weight.
Why grooming skips the cards that need it most
A grooming session naturally gravitates to what’s new, because new cards are quick to judge — the context is fresh. Old cards are the opposite: judging whether one still matters means reconstructing why it was added in the first place, which is slow, so it gets deferred to next time, indefinitely.
The backlog grows lopsided as a result — a thin layer of recently-groomed cards sitting on top of a thick sediment of untouched ones, most of which nobody could currently justify keeping if actually asked.
What a measured groom returns
Say you run product for veterinary-clinic scheduling software and have fixed the standard that decides this for you: a card stays in the backlog only if it still serves a live objective, not merely because it was once a good idea. Nine cards have sat untouched for two quarters or more.
Measured against your own standard, the groom returns a specific list, not a general tidy-up: “Nine stale cards serve no live objective — your archive list.” None had been reviewed since being added. All nine had quietly outlived whatever objective justified them originally.
What grooming becomes
A generic staleness filter can flag cards by age, but age alone isn’t your test — a two-year-old card tied to a current objective should stay, and a two-month-old card tied to nothing should go, which needs your live-objective rule applied, not a simple date filter.
Measured context checks every card, old and new, against your live-objective standard directly, so grooming stops being a skim of the top of the list and returns exactly which cards are dead weight regardless of how long they’ve been sitting there.
Backlog grooming naturally skips the old cards that need judging most, because reconstructing why they were added is slow work; through Unl the PM’s own live-objective standard checks every card, old and new, and names exactly what’s dead weight.
Reads through Unl arrive with measured context — in the presence of the decisions you’ve already settled. It works today: one box, paste anything. If it speaks MCP, Unl can reach it. Readings arrive unprompted, the data beside the criterion; Unl is a courier, not a warehouse, and keeps only your keys and the frame.
Read further
Questions people ask
Why does backlog grooming never actually clear out old cards?
Because judging whether an old card still matters means reconstructing why it was added in the first place, which is slow, so grooming naturally gravitates to the newer, quicker-to-judge cards instead. The backlog grows lopsided as a result — recently-groomed on top, untouched sediment underneath.
What’s the right test for whether a backlog card should stay?
Whether it still serves a live objective, not how old it is or how good an idea it once was. For one vertical-SaaS PM, that’s the whole rule — a card with no live objective behind it gets archived, regardless of when it was added or how it looked at the time.
Can AI groom my backlog for me?
It can flag cards by age, but age alone isn’t the right test — a card tied to a current objective should stay however old it is. Measured context applies your live-objective standard to every card directly, old and new alike. It works today: one box, paste anything. If it speaks MCP, Unl can reach it. Readings arrive unprompted, the data beside the criterion; Unl is a courier, not a warehouse, and keeps only your keys and the frame.
What if nobody remembers which objective an old card was meant to serve?
Then it cannot be matched to a live objective, and the read names it as unlinked. A card nobody can connect to a current goal is usually a card to cut, and the decision to keep it anyway is yours.
What this is
Your agents don’t need to persist. Your purpose does.
Save the thoughts, decisions and targets worth keeping, each with its reasoning, carried into every AI session the moment they matter. A new unit of exchange between you and your AI: the Settled Why with standing that travels. Unprompted.
Your whole AI world. What you decided at the epicentre. It works inside Claude, Claude Code, ChatGPT and Cursor, and connects in a couple of steps.
MCP native·Human settled·Model agnostic·Your data
Measured Context
Connect Unl to bring the right information into the moment.
Your sources, read against the criteria you set.
The full product, open. Free to start. Paid plans from $10 a month.