What it is
Data lineage is the traceable history of a single piece of data: where it originated, what it passed through and what changed along the way. For a recruitment desk that usually means one field in a CRM or ATS record, a decision-maker's email, a headcount figure, a "hiring" flag, and the chain behind it: which provider or signal first surfaced it, which enrichment pass verified or corrected it, and whether a consultant edited it by hand before it reached a task. Lineage is not the value itself. It is the path the value took to get there.
It is a close cousin of data provenance, which records who or what first created a piece of data and under what authority, and it sits next to an AI audit trail, which logs what a system did. Lineage is narrower than both: it follows one data point through every transformation it went through, not the whole system's activity and not just its point of origin. A verified email has provenance, whichever provider first supplied it, and lineage: that provenance, plus every waterfall step, edit and re-check that happened after.
A field with no known source is a guess wearing a CRM label.
Why it matters
Recruitment data decays and gets stitched together from many sources, which means most CRM records are already a patchwork before anyone edits them by hand. Without lineage, a wrong number is just wrong, with no way to tell whether it came from a stale provider, a guessed email format or a typo a consultant made six months ago. With lineage, a consultant can see exactly which link in the chain to fix, and whether the same source is quietly wrong across other records too.
It also settles disputes that would otherwise become guesswork: two enrichment passes disagree on a company's headcount, or a contact's title looks out of date. Lineage tells you which source is more recent and which pass actually verified the field, rather than which one happened to load first. That matters even more once AI is doing the enrichment, because a system correcting its own earlier guess needs a record of the guess it is correcting.
How boilr handles it
In boilr, every field your AI sales employee writes into an account, a verified email, a headcount, a signal it detected, carries its lineage with it inside the Company Brain: which source supplied it, which step in the enrichment waterfall confirmed it and the timestamp it was last checked. You are never looking at a bare number with no way to ask where it came from.
That lineage is what makes the Company Brain trustworthy rather than just large. When a field changes, the record shows what replaced what and why, so a consultant reviewing a task can trust an enriched detail instead of re-checking it by hand, and your agency can show a client or a regulator exactly how a specific fact in your system came to be there.
