What Should a Commercial HUMINT Report Contain?

Start with the decision and the precise questions

A commercial HUMINT report should make the assignment understandable to someone who did not commission it. State the decision, companies or commercial relationships in scope, relevant period and questions being answered. A buyer assessing a supplier's operating reputation needs a different account from counsel testing a disputed introduction. A collection of opinions about a business is not a substitute for a defined question.

The executive summary should give the material answers, their significance and the unresolved issues. It should say whether the assessment concerns historical conduct, a current position or expectations about future behaviour. Dates matter when a source's knowledge comes from a former role or an earlier commercial relationship.

The public PHIA analytical standards offer useful principles on relevance and traceable judgements. These are government assessment benchmarks that a client can adapt when specifying a deliverable, not certification of a private provider or a mandatory template for every commercial assignment.

Ask the provider to identify the material supplied by the client and any important exclusions. A conclusion should not appear comprehensive if critical questions were outside the mandate or depended on documents that were never available.

Make the basis of human knowledge reviewable

A source descriptor should explain why the account is relevant without unnecessarily exposing the person's identity. Useful context includes the person's relationship to the subject, the period of their knowledge, whether they observed the event directly and what limits their view. A prestigious job title alone does not establish knowledge of the particular transaction.

These descriptors help assess weight; they do not require a client-facing report to publish private identifying details. Agree how the reviewer can understand the basis of an account, who may see supporting material and what confidence limits remain when identities or underlying records cannot be disclosed.

Our source-validation guide explains the distinction between assessing a person's relevance and testing the particular statement attributed to them.

Separate accounts, corroboration and analytical judgement

Use language that preserves the evidence step. “A source described repeated approval delays” records an account. “The delays occurred in every case” is a broader claim requiring additional support. “Management may have prioritised another customer” is an interpretation that needs reasoning and consideration of alternatives.

Where personal information is involved, the ICO's accuracy guidance stresses clarity about source and status, including attributed opinions. A commercial report should therefore avoid turning an opinion about an individual into an established finding merely by removing its attribution from the summary.

The PHIA uncertainty framework distinguishes assessed likelihood from confidence in the basis of an assessment. Ask the provider to explain its own terminology and the actual limits: direct but historical knowledge, conflicting accounts, narrow source coverage or missing corroboration.

An unexplained score or colour cannot replace that explanation. Confidence should attach to the material assessment, not automatically to every statement made by a source who was useful on another topic. Contrary information belongs alongside the finding it affects.

Read a clearly fictional qualified finding

Fictional illustration: a client is considering an agreement with invented Supplier A. Two simulated source accounts describe slow approvals during a past project. The people, company and records in this example are fictional and are not a completed provider case.

Account S1 describes approval meetings the fictional source attended during the review period. Account S2 repeats a colleague's description of the same project. A simulated client record confirms one delayed approval but says nothing about the cause. The evidence supports a question about that project's approvals; it does not establish a current company-wide pattern or deliberate obstruction.

A useful finding would identify S1's direct proximity, S2's possible dependence on the same story, the record's narrow corroboration and the unresolved explanation. It might recommend requesting the relevant project chronology before using the issue in the partnership decision.

For corporate HUMINT assignments, this is a practical output: qualified context that helps the client choose the next evidence question. The report should not convert limited historical accounts into a confident present-day accusation.

Keep qualifications when the report is shortened or shared

A decision brief, full findings and supporting annex can serve different readers. Agree which version goes to management, counsel or another authorised adviser. A shortened version should retain the date, scope and material confidence limits rather than preserving only the strongest-sounding sentence.

If identifying or sensitive details are restricted, make the restriction understandable. State what source context can be disclosed and what it prevents the reader from independently checking. A confidentiality label does not itself establish legal privilege, credibility or permission to share personal information.

The HUMINT confidentiality guide addresses handling arrangements. For report review, the central question is whether the reader still sees the same qualified conclusion after redaction, translation or summarisation.

Use version control. A later correction should identify the changed source account or record, why it changed and whether earlier assessments still hold. Do not silently replace a document on which a committee may already have relied, leaving the recipients unaware of a material change.

Review the evidence trail before accepting the deliverable

Use a short acceptance checklist:

• The report answers the agreed questions and identifies exclusions.
• Source knowledge is relevant to the claimed event and period.
• Direct accounts, reported information and interpretation are separated.
• Repeated accounts are not presented as independent confirmation without a basis.
• Supporting records, contrary information and gaps remain visible.
• Confidence statements explain what could change the assessment.
• Sharing versions preserve the material qualifications and update history.

When comparing best HUMINT service providers, ask for a redacted example that demonstrates these features. Discuss how source protection affects the proposed evidence explanation rather than accepting secrecy as a substitute for reviewable reasoning.

The purpose of a good report is to improve the decision, including identifying an allegation that cannot be supported. A candid gap or a finding that contradicts the starting theory can be more useful than a polished account that overstates what people actually knew.