Executive Summary: Multi-state business search is a candidate-generation problem first and an evidence-based entity-resolution problem second. The same legal business may have a domestic formation record and foreign registrations in other states. Different businesses may also share a similar name. A safe workflow therefore collects state candidates broadly, then clusters records only when formation details, identifiers, addresses, people, and other evidence support the relationship.
Why Can One Business Have Several State Records?
SBA guidance distinguishes the state where a business is formed from foreign qualification in other states where it conducts regulated activity.[1] A Delaware corporation registered in California and Texas can appear in three state systems while remaining one legal entity.
That does not mean every same-name result is the same company. State databases differ in naming, identifiers, fields, and availability.
A state hit is a candidate until evidence connects it to the legal subject.
What Inputs Should the Search Start With?
Preserve the application values and build a query set:
1. Exact legal name.
2. DBA and known aliases.
3. Formation state.
4. Known foreign-registration states.
5. State entity IDs.
6. Principal and mailing addresses.
7. Registered agent.
8. Officers or managers where published.
9. Prior legal names.
Do not treat a shared registered agent as strong identity evidence by itself. Agents represent many unrelated companies.
How Does Cobalt Support the Search?
Cobalt's documented SOS product supports single-state searches using available inputs such as business name, state, entity ID, person name, or address. Full Verification searches 50 states plus DC asynchronously when the state is unknown and returns state-organized matches and alternatives.[2]
Full Verification discovers candidate state records. It does not prove beneficial ownership, corporate hierarchy, or that every similar result is one business.
subject identity
-> multi-state candidate collection
-> record-level evidence comparison
-> cluster, separate, or review
Which Evidence Can Link State Records?
Use a weighted evidence set rather than one field.
• Formation jurisdiction and formation date.
• Domestic entity identifier referenced by a foreign record.
• Exact legal name and known prior names.
• Principal address history.
• Consistent officers or managers where available.
• Official foreign-registration relationship.
• Filing documents or certificates when required.
California's business-search definitions explain that a foreign entity's status in California concerns its California registration, not necessarily its standing in the formation state.[3] Preserve the jurisdiction-specific meaning.
What Evidence Is Weak by Itself?
• Similar display name.
• Same generic address.
• Same commercial registered agent.
• Same industry.
• Same website branding.
• A name token shared by a corporate family.
Weak evidence can help generate a candidate but should not complete a cluster.
How Should the Resolution States Work?
• Confirmed same entity. Strong official or multi-field evidence links the records.
• Likely same, review required. Several fields agree, but a key identifier is absent.
• Separate legal entities. Evidence identifies distinct formations or tax identities.
• Related but not the same. Parent, subsidiary, franchise, or affiliate relationship is documented.
• Unresolved candidate. Evidence is insufficient.
• Source unavailable. The state could not be searched reliably.
Do not force unresolved records into a cluster to simplify the UI.
How Should Foreign Registrations Be Represented?
Keep one canonical legal-entity record with linked jurisdictional registrations. Each state record retains its own identifier, status, source value, retrieval time, and evidence.
This design lets policy ask the correct question. The lender may need to know whether the entity exists in its formation state, whether it is registered in the operating state, or whether a status issue exists in one specific jurisdiction.
What Should Happen When Names Change?
Preserve the prior name, current name, filing evidence, and effective date where available. Search both names under the relevant state's rules. A new name should create a new identity version, not erase the historical record.
If a same-name record has a different formation state, date, or entity ID, keep it separate until stronger evidence proves a relationship.
What Should the Audit Trail Retain?
• Original subject inputs.
• Every state searched and its coverage state.
• Exact queries.
• All candidate records.
• Source timestamps and availability.
• Evidence fields used for each link.
• Resolution state and reviewer.
• Policy version and downstream use.
For using these events in a credit record, see How Risk and Underwriting Teams Should Use Automated Audit Trails.
What Should Teams Monitor?
1. Candidates per search.
2. Records automatically clustered.
3. Clusters reopened by reviewers.
4. Foreign registrations linked to a formation record.
5. Same-name false matches.
6. States returning unavailable or incomplete evidence.
7. Searches requiring prior-name variants.
8. Decisions made while identity remained unresolved.
These measures test entity resolution. They do not establish ownership or creditworthiness.
What Is the Practical Multi-State Sequence?
Collect candidates across relevant states, preserve each record independently, compare strong identity evidence, classify the relationship, and route ambiguity to review. Only confirmed clusters should feed a consolidated customer record.
Cobalt can supply single-state and 50-state-plus-DC search results. The lender owns clustering rules, relationship evidence, policy, and decisions.
Ready to test one foreign-qualified entity, one same-name false match, and one unresolved candidate? Schedule a Cobalt Intelligence demo and bring the identifiers your current customer master retains.












.png)