What the broker didn't send

A submission tells you what the insured knows, or what the broker could extract from the insured on a Tuesday. The gaps are rarely deliberate.

Square footage goes missing, the parcel is described rather than identified, the valuation is whatever was on last year’s schedule, and the location is an address that may or may not geocode to the building you are being asked to insure. Atlas fills those gaps from aggregated third-party sources, and tells you where every value came from.

Why underwriters ignore enrichment

Because the objection is never answered. Enriched data arrives as a confident figure with no attribution, and an underwriter who is accountable for the bind has no way to work out whether it came from an authoritative register or from something scraped and resold. The second question follows immediately: how old is it? A valuation that was accurate four renewals ago is not a valuation, it is a liability with a decimal point.

So the data gets added, nobody trusts it, and it does no work.

Atlas is built the other way round. Attribution is not a report you can run afterwards. It is in the field.

What an Atlas field carries

Every enriched value returns with three things attached to it: the source it came from, an indication of data quality, and the age of the data. Not at the record level. Per field.

The structure, illustratively. Field names are indicative and no values are shown, because a sample value on a web page is a number someone will quote back at us.

field
  name                  // e.g. gross building area
  value
  source
    name                // the originating data source
    reference           // identifier or record within it
  quality               // quality indication for this value
  as_of                 // the date the source data applies to

An underwriter looking at an Atlas-enriched record can see, on the field itself, that the square footage came from one source and is recent while the valuation came from another and is not. That changes the decision from “do I trust enrichment” to “do I trust this value”, which is a question an underwriter can actually answer. Fields returned include size, location, parcel, exposure, valuation and price per square foot, among others.

One field, returned by Atlas

gross_building_areavalue

never returned on its own

sourcewhere it came fromNational parcel registerrecord APN 118-204-0037
qualityhow good that source is hereVerified against registerauthoritative, not derived
as_ofhow old the data isMarch 2026the date the source data applies to, not the date it was fetched

Per field, not per record. The question stops being “do I trust enrichment” and becomes “do I trust this value”.

The anatomy of one enriched field. The field name, source name and date are illustrative and fictional. No value is shown: the point of the figure is everything attached to it.

It is the provenance argument again

SnapLine's position on extracted data is that every field traces back to the region of the document it was read from. Atlas applies the same rule to data the broker never sent. The underwriter is not asked to accept a value because the system produced it, whether the system read it off a schedule or fetched it from a register.

One discipline, two problems. If you have already seen the provenance view, you know what this is.

Licensing

Atlas is licensed separately from SnapLine rather than bundled into it. If you are evaluating both, say so when you book and the call covers both scopes.

Bring a submission with holes in it

Not your cleanest one. The one where half the schedule is blank and the broker has stopped replying.

Up to three of your own submissions can go into the demo alongside the pre-loaded samples. What is worth watching is not that Atlas fills the gaps. It is what it says about each thing it filled.

Book a demo