Think inside your AI world.
BigQuery + PostHog through Unl
Your warehouse and your product analytics can each report “retention” — and mean different things. The definition that reconciles them isn’t in either.
BigQuery + PostHog through Unl reads a warehouse cohort and a product funnel together against the single retention definition you ratified, so both answer to one meaning — agreeing, or disagreeing with the reason — not two numbers that look alike.
The criterion that binds them
You ratified one definition of retention — the window, the qualifying action, the denominator — because a warehouse query and a product funnel will each invent their own otherwise, and the two will quietly disagree.
The two naked reads
BigQuery returns a cohort table; PostHog returns a funnel. Both accurate on their own terms, and their terms differ — so “what’s our retention?” has two answers and no arbiter.
The one measured answer
Measured against your definition: on the retention rule you ratified, both sources land at 24% — they agree; the earlier 31% from the warehouse used a looser window than your definition allows. One definition over both reads turns two numbers into one trusted answer.
And back again
When you tighten the qualifying action in the retention definition, that change is ratified once in Unl — and both the BigQuery and PostHog reads move to the new meaning together.
The answer comes back measured against what you already decided, and why.
A router can run a BigQuery query and pull a PostHog funnel. It cannot make them mean the same thing, because the retention definition — the criterion — lives in neither warehouse nor product tool.
The lane is live and open to this tool 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
Can AI reconcile retention across BigQuery and PostHog?
Yes, through Unl. You ratify one retention definition — window, action, denominator — and the BigQuery cohort and PostHog funnel are read against it, so both answer to the same meaning and any disagreement is explained.
Why do my two tools report different retention?
Because each applies its own default definition. The single definition that reconciles them lives in Unl, so a measured read holds both to it.
Does Unl store my warehouse or product data?
Unl reads through your warehouse and product data, and can write back on your explicit gesture — it never acts as a side effect of a read.
What if the warehouse and the product disagree on when a day ends?
Then the two reads are answering the same definition on different clocks, and the retention figures will diverge without either being wrong. A timezone boundary is exactly the kind of quiet disagreement one ratified definition is meant to remove, and it is not removed by naming the window alone. The definition has to fix the clock as well, or the two systems will keep inventing their own.
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.