What match rate can we expect?

On comparable inputs, customers typically see match rates climb from the 30s to the 70s after switching. A top-10 Canadian carrier moved from 35% to 75% on the same dataset replacing a legacy bureau.

How are subsidiaries and parent companies handled?

Resolution maintains the full corporate hierarchy: legal parents, operating subsidiaries, alias chains, and historical identities. A sanctions hit on a parent fires across linked subsidiaries; revenue rolls up to the holding company without losing the operating identity.

Does resolution work outside English?

Yes. Arabic, Chinese, Cyrillic, and Hebrew scripts resolve at the same accuracy as Latin scripts. Cross-script aliases (the same company spelled four ways across four languages) collapse to one canonical entity.

Can it run on our existing data?

Yes. Resolution accepts your records and returns resolved IDs you can join back into your warehouse.

Services / Entity Resolution

One record per company, no matter the source

Most tools treat a company as its legal registration. Veridion resolves the full picture: legal entity fused with digital presence, relationships, and subsidiaries, into one verified, defensible profile.

99.2% resolved on first pass · 42,108 records · 3m 14s.

Discuss a project View company enrichment

Before5 raw inputs

STRIPE INC
Stripe Payments Co.
stripe.com llc
Stripe (San Francisco)
Stripe Pmts

After1 canonical record

SI

LEI: 549300CLHGIPTCYHQ143
Domain: stripe.com
Sector: Payments & Fintech
Confidence: 0.97

Failure modes

Where resolution fails.

Without a resolution layer, this is what breaks in your own systems.

Duplicates

What happens when the same company appears four ways across CRM, ERP, and billing?

Every report, model, and rollup built on top is quietly wrong.

Cross-jurisdiction

What happens when an entity exists in three registries under different legal identifiers?

They are separate legal entities, but they belong to one company. Your system leaves them unlinked, as three unrelated records.

Legal vs operating

What happens when a customer contracts under its parent company but pays under a subsidiary?

Your systems see two customers. Revenue, renewals, and account history split across records that never join.

Compliance risk

What happens when KYC and sanctions screening fail silently on unresolved entities?

A sanctions match that should fire on a subsidiary fails to fire, because the subsidiary is not linked.

507MLegal entities analyzed

135MValidated digital entities

7.8MCompanies updated in the last 30 days

How resolution works

Four steps from raw signal to canonical entity.

Crawl, recognize, fuse, filter. The graph updates continuously as new signals arrive, not on a quarterly batch cycle.

Collection
Web pages, registries, filings, and news. Crawled globally.
Entity Recognition
Our in-house extraction model processes names, addresses, IDs. Every language and writing system supported.
Graph Disambiguation
Confidence-weighted graph. One node per real-world entity, across languages.
Junk Filtering
400M sites analyzed, 100M verified. Spam and ghost entities are excluded; all valid entities are retained.

Why teams switch

Where legacy resolution stops, and we keep going.

Single-source fuzzy matching collapses on cross-jurisdiction, subsidiaries, and digital-only entities. We filter 400M sites down to 100M trusted ones before fusion, then refresh the graph continuously as new signals are crawled.

Legacy resolutionVeridion

Legal registryname + address

Fuzzy matchname + address onlyno confidence model

unresolvedno confident match

duplicatetwo IDs, one entity

unresolved · duplicateno confident match

Filings · footprint · graphthree layers fused

Legal filingsglobal registries

Digital footprintdomain, site, tech

Relationshipsparent, sub, alias graph

Knowledge graphweighted disambiguationspam and ghosts removed first

Canonical entityvrd_ent_8421conf 0.97

Canonical entityvrd_ent_2193conf 0.94

Canonical entityvrd_ent_7064conf 0.91

Canonical entitiesresolved with confidence

Single source · fuzzy match · ambiguous outputs.Three layers fused · continuous refresh · canonical IDs.

Live output

Four rows from a real resolution batch.

Canonical name, registry IDs, confidence, and signal flags for every resolved entity. Low-confidence records show explicitly, not hidden.

#	Canonical name	Domain	Confidence	Signals
01	Stripe, Inc.	stripe.com	x 0.98	domainlei
02	Shopify Inc.	shopify.com	x 0.94	nameaddr
03	Snowflake Computing Inc.	snowflake.com	x 0.91	parent
04	Datadog Holdings Sàrl	no domain resolved	x 0.42	low-conf

Full-batch distribution across 42,108 records>96% match rate is a floor, not an average.

1% 0.0 to 0.5
2% 0.5 to 0.7
7% 0.7 to 0.85
30% 0.85 to 0.95
60% 0.95 to 1.0

Batch · 42,108 records · 3m 14s · 41,762 resolved · 346 below thresholdTry the API

Proof

Where this has been applied

Tested directly against leading competitors in financial services and location intelligence, where accuracy is the difference between a good model and a broken one.

29,000 more businesses than the incumbent. 0 linkage errors against a 20% competitor error rate.

Experian · Australia

Match rate climbed from 35% to 75% on the same input dataset, replacing a legacy bureau.

Top-10 Canadian carrier · D&B replacement

Location-specific employee and revenue counts across 45 countries, in one canonical layer.

Global market intelligence platform

What you get

Deliverables

Resolution logic, match confidence, subsidiary linkage, and relationship graph data included.

Canonical entity IDs

Stable, system-agnostic identifiers for every resolved entity, with full resolution lineage attached.

Ships with

Resolution rules and assumptions
Lineage per entity

Match confidence and reasoning

Each resolution decision documented with the match reasoning and the signals that drove it. Auditable at the record level.

Ships with

Match reasoning
Per-record auditability
Validation samples
Edge-case documentation

Subsidiary and parent linkage

Entities linked to their corporate hierarchy, alias chains, and historical identities. Relationships captured from both legal filings and digital sources.

Ships with