Think inside your AI world.
Does the data meet our criteria?
Strip the questions people bring to their data and nearly all of them reduce to one: does this meet the criteria we set? “How’s conversion,” “are we growing,” “is churn fine” — each is a polite proxy for met or not-met, against our bar. Ask the real question directly and you discover the answer was never in the data alone; it was in the data and the bar, together.
“What is the number” is answerable by any tool. “Does the number meet our criteria” is answerable only where the criteria live. That is the whole distinction between description and judgement in a single question. Unl holds the criteria, so the question can be asked as itself — and the answer comes back met or not-met, naming the bar and why it is the bar, rather than a figure you have to finish in your head.
Why is this the real question?
Because the figure on its own settles nothing. “Conversion is 3.4%” is not good news or bad news until it meets a standard; it is a fact awaiting a comparison. Every question people actually ask their data carries that comparison implicitly — they do not want the number for its own sake, they want to know whether it clears the line they care about. The number is a means; meets our criteria is the end.
Asking it directly is clarifying, because it names the two ingredients the answer needs: the read (from the tool) and the criterion (from you). It also exposes why so many analytics answers feel unsatisfying — they deliver the first ingredient and quietly omit the second, handing you a fact and calling it an answer when the comparison you actually wanted was left undone.
Why can only the frame answer it?
Because half the question lives on your side. Say you lead product and have ratified a clear criterion for a feature launch: a new feature stays in the product only if week-four retention of its adopters clears 30%, because below that it is adding surface area without adding stickiness. Any tool can tell you adopter retention is 26%. None can tell you that fails, because the 30% bar and its reasoning are not data the tool holds — they are a decision you made.
So “does this meet our criteria” is unanswerable by whatever renders the metric, and answerable only where the criterion is held. The read comes back as “week-four retention of adopters is 26%, under the 30% keep-it line you set, so this feature isn’t earning its surface area — that’s the reason the line is 30%.” Met-or-not, with the bar and its why, which is the shape of an answer rather than a fact.
What does asking it directly change?
It changes the default output of analytics from figure to verdict, and it changes what a review meeting is for. If the question “does this meet our criteria” is answered at read time, the meeting no longer exists to run the comparison — the comparison is already done, against ratified bars, on record. What is left for the room is the genuinely human part: deciding whether the criteria themselves still hold, and changing them when they don’t.
That is the destination the whole field spine points at. Description was never the goal; it was the raw material. The question that mattered — does this meet the criteria we set — needed a layer to hold the criteria, and once that layer exists the question can finally be asked plainly and answered as a verdict. The rest was always just fetching the number the verdict is built from.
Nearly every analytics question is a proxy for one real question — does this meet the criteria we set? — and its answer needs both the read, which any tool supplies, and the ratified bar, which only you hold, so a verdict is sayable only where the criteria live.
Reads through Unl arrive with measured context — in the presence of the decisions you’ve already settled. The reach lane is live: 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
What’s the real question behind ‘how are our numbers’?
It’s ‘does this meet the criteria we set?’ A figure on its own settles nothing — ‘conversion is 3.4%’ is a fact awaiting a comparison, not good or bad news yet. People want the number only to know whether it clears the line they care about, so the operative question is met-or-not against your bar, not what the number is.
Why can’t my analytics tool answer whether we’re meeting our criteria?
Because half the question lives on your side. The tool can say adopter retention is 26%; it can’t say that fails, because the 30% keep-it bar and the reason behind it are a decision you made, not data the tool holds. ‘Meets our criteria’ is answerable only where the criteria are held, not where the metric is rendered.
What does answering it at read time change?
It turns the default output of analytics from a figure into a verdict, and it changes what a review meeting is for — the comparison is already done against ratified bars, so the room is left with the human part: deciding whether the criteria still hold. The reach lane is live: 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 we haven’t actually agreed our criteria yet?
That is the common case, and finding it out is worth more than any read you were about to take. Most teams have criteria in practice — carried privately, applied inconsistently, and discovered only when two people read the same number differently. Say one of you treats a fortnight of flat signups as noise and another treats it as the signal to change the plan: that gap was always there, the number simply exposed it. The first pass is not instrumentation, it is agreeing what the bar is and why it sits there, and everything measured afterwards answers to something real.
What this is
Think inside your AI world — you stay in command
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 plugs into Claude, Claude Code, ChatGPT and Cursor as an MCP connector — quick to connect, 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 at launch.