The disciplined path every opinion follows — and the rules that hold it in place when convenience pushes back. Version 2, September 2026.
Every audit object arrives with two layers: what is claimed about it and what is proven about it. The methodology exists to measure the distance between the two — and to state it clearly, before the mistake becomes costly. Intentions, projections or narratives are not validated. Evidence is.
Not every piece of evidence carries the same weight. The opinion states, for each finding, which rung the evidence rests on:
A finding supported only by the last two rungs does not come out as verified.
No statement is left without a state. A finding is Verified when, and only when, the evidence from the first three rungs supports the claim, the trail is recorded and a third party would retrace the path. It is Assumed in any other case: claimed and not yet exercised; supported only by a document from the party or by a statement; or verified under a condition different from the one claimed (in the demonstration, not in production; in the past harvest, not the declared one). The assumed is not discarded. It is named — and the opinion states what is missing to cross the boundary.
Crossing the boundary is a recorded act: date, new evidence, who examined it. Nothing crosses out of inertia, deadline, or request.
The state says whether the claim holds up. The grade says how much it matters:
Grade is assigned by two examiners, and any disagreement is recorded when it exists.
Auditing everything is rarely possible; pretending everything was audited is the most common failure. The scope letter declares coverage: how many items exist, how many were examined, and how they were chosen. When there is a sample, it is risk-directed — the items of greatest value, of greatest historical deviation, or flagged by a signal — and the opinion states, in numbers, what was left out. What was left out is neither assumed nor verified. It is out of scope, and that is how it is written.
Every finding occupies one line, with the same columns: number · claim as it arrived · state · grade · evidence (rung, origin, date, examiner) · caveat · recommendation · what is missing to verify. The matrix is the opinion in table form; the opinion text is the matrix in reading form. The two say the same thing, and the matrix prevails in case of doubt. Template at /modelos/matriz-de-achados.
A caveat is not an excuse; it is a declared limit. The four families:
A caveat with no family does not enter the opinion.
No opinion goes out with just one pair of eyes. A second examiner, who did not take part in the collection, redoes the evidence trail by sample — and answers the only question that matters: with the same base, would they arrive at the same conclusion? If the answer is not a comfortable yes, the opinion goes back. It is this test, not the signature, that makes a document an audit.
Independence is practiced through rules, not adjectives. Before accepting an object, we check for conflict: a corporate, commercial or personal tie with whoever hires us or whoever is audited. An existing tie is declared on the first page of the opinion; a tie that compromises judgment blocks the engagement. The fee is set with the scope, before starting, and does not depend on the conclusion — nor on a deal's closing timeline, nor on a transaction's outcome. When the object belongs to the group itself, the same ruler applies, and the tie is stated. The full letter is on the home page, at Letter of independence.
The opinion belongs to whoever commissioned it. Valida does not display clients or their numbers — not as a case, nor as a statistic. Documents received are processed in accordance with the LGPD, kept on servers in Brazil, and returned or deleted at the end of the retention period agreed in the scope letter, with a record. Sensitive data — health, payroll, biometrics — has its own access trail and is examined by the negative: whoever should not see it, we try to make them see it.
The evidence dossier is kept for the period set out in the letter, with an index and a cryptographic hash of each file, so that any party can later prove that what was examined is what is kept. Reproducing the opinion is the right of whoever commissioned it: the base, the criteria and the path are in the appendix, and a competent third party should arrive at the same place.
It is not an accounting audit, nor an opinion on financial statements. It is not a valuation report. It is not a legal opinion. It is not a certification on behalf of any standard. It is the answer, substantiated and reproducible, to a bounded question: does what is claimed about this object withstand the evidence?
Three documents, published as templates so that no one needs to imagine what the result will look like: the scope letter, the findings matrix and the opinion. Illustrative examples, with no real development. At /modelos.
Every finding comes out with the same design: what was claimed, what the evidence supports, what still lies at rest — and the trail that allows a third party to retrace the path.
By the boundary, written into the scope letter: the relationship is declared on the first page, the fee does not depend on the conclusion, and the evidence prevails over the interest of whoever commissioned it. The opinion on the group follows the same yardstick as the opinion on third parties — and the third-party test applies to both.
No. It is technical audit and fact validation: what is claimed about a development, a harvest, a resource, a platform, a team or a protocol, confronted with the evidence. Financial statements have their own auditor; when the opinion touches on numbers, it treats them as a claim to verify, not as an accounting opinion.
What was claimed and not yet exercised. It is not a judgment: it is a state. The assumed stays named in the opinion so no one confuses it with verified — and so it is known exactly what is missing to cross the boundary.
The scope letter arrives within three business days after the initial conversation. The opinion's timeline is defined in it, in writing, according to the object — and it is the timeline that is kept.
The fee is set with the scope, before starting, and does not change with the conclusion. There is no proposal without a scope: a price with no bounded object would be the work's first unverified claim.
The signed opinion, with each claim classified, the evidence trail attached, the caveats explicit and the recommendations. And a reading meeting.
A named technical lead, with the relationship declared. An opinion without a name is not an opinion.
Yes. Full confidentiality, processing in accordance with the LGPD, data stored on servers in Brazil. The Privacy Policy names each operator.
Yes. The scope can be narrow, as long as it is closed. What is not in the scope is stated as out of it — never presumed.
It prevails. It is in the letter, before the signature.