Think inside your AI world.
Analytics only ever shipped description
Every generation of analytics — the report, the dashboard, the warehouse, the semantic layer — got better at answering “what happened.” None of them answered “is that good enough, and why.” That second question was left, silently, as an exercise for the reader. The field shipped description and called it insight.
The semantic layer was a real achievement: it made revenue mean the same thing in every chart, so teams stopped arguing about definitions. But standardising what a number means is not the same as holding the standard it must meet. Unl is the layer the field skipped — the criteria layer, where the bar for each number, and the reason the bar sits there, are ratified and travel with the read.
What has analytics actually been solving for?
Trace the lineage and it is one long improvement in description. Reports made last quarter legible. Dashboards made it live. The warehouse made it joinable. The semantic layer made active user and net revenue mean one thing across the company. Each step was a genuine advance, and each stayed on the same side of the line: telling you what is, more accurately and more quickly.
The other question — whether what is, is acceptable — never got a layer of its own. It was handled the way it always had been: a person looked at the number, recalled the standard they held in their head, and decided. The tooling got faster underneath a judgement that never got any help at all.
Why did the semantic layer not close the gap?
Because a semantic layer standardises meaning, not standards. It guarantees that when two dashboards say “churn” they mean the identical calculation — a real and hard-won thing. It says nothing about whether a given churn figure is tolerable, because tolerability is not a definition; it is a position someone took, with a reason, that can differ by company and change by quarter.
Say you are a growth lead. Your semantic layer resolves “7-day retention” perfectly and identically for everyone. It cannot tell you that 34% is a miss, because the line you care about — 42%, below which the referral loop stops compounding — is not a definition to standardise. It is a criterion you ratified. The layer that agrees on what the number is has nothing to say about whether it is enough.
What does the missing layer look like?
The criteria layer holds, per number, the bar it must clear and the reasoning that put the bar there — and applies it at read time. For you that means “7-day retention is 34% against the 42% you set for the referral loop; below the line, so the loop won’t compound this cohort.” The semantic layer told you what retention is; the criteria layer tells you whether it passes, and why passing was defined that way.
This is not a smarter chart or a better summary of the chart. A summary still describes. The criteria layer changes the output type from figure to verdict — and because the verdict names the ratified bar and its reason, it is a judgement you can act on and audit, not one you have to reconstruct from a number and a memory.
Analytics has always shipped description — what happened, ever faster — and left ‘is that good enough, and why’ to the reader; the criteria layer is the missing piece that holds each number’s ratified bar and returns a verdict instead of a figure.
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 has analytics never solved?
It has never shipped judgement. Reports, dashboards, warehouses and the semantic layer all got better at describing what happened; none of them hold the standard a number must meet. That standard — the bar and the reason it sits there — was always left to a person to supply from memory, which is the gap the criteria layer fills.
Isn’t the semantic layer already the answer?
The semantic layer standardises what a number means, so “retention” is the same calculation everywhere — genuinely valuable, and not the same thing as holding the standard the number must meet. Whether 34% retention is acceptable is a position someone took with a reason, not a definition to standardise. That position is what the criteria layer ratifies and applies.
What is the criteria layer?
It is the layer that holds, for each number, the bar it must clear and the reasoning behind the bar, and applies that at read time — so “retention is 34%” comes back as “34% against the 42% you set for the referral loop, a miss.” It turns description into a verdict you can act on and audit. 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 this is
Think inside your AI world — you stay in command
Unlimitless (Unl to friends) holds what you've settled, reads what your tools are showing, and catches what's changed out in the world — and hands your AI whatever bears on the work, the moment it's needed, without you asking. The right thing, in front of the model, unprompted, with you in command of the call. So you keep moving toward what you set out to build, on top of everything you've already decided.
It plugs into Claude, Claude Code, ChatGPT and Cursor as an MCP connector. Quick to connect, in a couple of steps.
Unlimitless is open now to invited Alpha. Apply for the Beta waitlist to come in ahead of the full launch:
Alpha is invite-only · free at launch.