D04
Read the result
Candidate documentation for reading and synthetic examples. Account creation, live API access, SDKs, collections and MCP are not available here. A released service profile and matching integration instructions are required before use.
Read three things separately: the business outcome, the information and evidence, and the usage receipt. An HTTP success does not by itself mean the company was identified.
data.outcome |
Meaning and next step |
|---|---|
established |
The request met the profile’s identity-establishment rule. Inspect the returned entity, explanation and limitations before using the result in your workflow. |
ambiguous |
More than one defensible candidate remains. Review the permitted candidates and supplied recovery guidance. Choosing a candidate in your interface does not improve the evidence. |
insufficient_evidence |
Available permitted evidence cannot establish the intended identity. Supply known additional information only where suggested, or stop when no useful next input is available. |
no_match |
No match was found within the completed requested scope. This does not prove that the legal entity does not exist elsewhere. Review scope and recovery guidance. |
The response envelope contains data, receipt and meta. Keep the original machine response intact. A readable summary is a view of that response, not a replacement source.
For an established result, data.entity.subject.entity_id is the ENTITY identity reference. Record the accompanying context and publication/receipt references needed for permitted follow-on use. An ENTITY identifier is distinct from a source registration number and does not collapse a parent and subsidiary into one legal entity. Check lifecycle information when an identity is corrected, split, superseded or otherwise changes; do not silently overwrite your records.
data.identity_assessment concerns whether the supplied reference identifies the intended company. data.entity.confidence_summary and the assessments attached to returned claims concern the supported information. Strong information about one candidate does not establish that it is your intended company.
When an assessment is available, show its percentage, named zone, explanation, scope, method and limitations together. The evidence index is not a calibrated probability, measured overall accuracy or a recommendation to approve a customer, supplier, insurance risk or transaction. “Not assessed” is not zero confidence and must not become a fabricated percentage.
The result may also explain normalisation and recovery. Do not infer that a normalisation explanation is a complete cleansed copy of your input. In a supplied/result/comparison view, a difference is something to examine; it is not automatically an error or an instruction to change your source record.
Worked example — read a synthetic result
Open the labelled example for HARBOUR LANE LOGISTICS LIMITED. This is fictional review data. It demonstrates how to read a response; it is not a lookup of a real company, and opening or copying it does not call the live service or spend credits.
Read the response in this order:
| What to inspect | What this example shows | What it means |
|---|---|---|
| Outcome | data.outcome is established |
This fixture identifies one company. Read the accompanying explanation and supplied context before treating an established result as the company you intended. |
| Company reference | data.entity.subject.entity_id is 7f3c9a1e-0000-4000-8000-0000000e0001 |
This is the example’s ENTITY identity reference. It is distinct from its sample registration number, 09876543. Preserve leading zeroes in registration identifiers. |
| Legal name | The legal_name claim contains HARBOUR LANE LOGISTICS LIMITED |
The name is returned as a claim with its own supporting evidence and limitations, not as an unsupported label. |
| Evidence | The legal-name claim’s evidence level is discovery; its explanation says there is no real source or verification |
An established identity outcome does not turn every attribute into an independently verified fact. Read each claim’s evidence scope. |
| Confidence | The legal-name claim’s delivered-support assessment is not_assessed, with no percentage |
“Not assessed” means there is no usable assessment here. It does not mean zero confidence or a fabricated confidence score. |
| Freshness | The legal-name claim’s freshness state is unknown |
Do not substitute the response time for the date a source fact became true or was verified. |
| Usage | receipt.settlement is settled, with charged_units equal to the string "1" and unit test_credit |
This is an illustrative receipt within the fixture, not a charge caused by viewing it or a published service tariff. In a real response, interpret the amount with the applicable offer, unit scale and settlement state. |
| Example status | meta.synthetic is true |
Keep the example label visible. This response is not evidence of a live lookup or released source coverage. |
Now compare the other three labelled examples. An ambiguous outcome requires you to inspect the candidates; it does not choose one on your behalf. No match means the request did not establish a matching company within its stated scope. Insufficient evidence means the available evidence cannot establish the intended identity. Neither outcome proves that a company does not exist. Follow the actual explanation and permitted recovery guidance instead of repeatedly submitting the same lookup.
Keep the original JSON available alongside the readable summary. When ready for a real request, use the supported signed-in tester or released integration instructions and review the applicable scope and maximum usage first.
This HARBOUR LANE reading exercise is separate from the Example Demonstration Limited request and response in the first-request guide. Neither synthetic example demonstrates an authenticated service exchange.
The offline TypeScript, Python and Elixir readers inspect the same unchanged synthetic responses, including receipt units, assessment context and recovery. They are response-reading examples, not released clients.
Inspect the complete four-outcome examples and readers · Open the unchanged outcome JSON