Translation Quality Verification: Why It Matters

A translation that arrives on time and looks properly formatted creates the impression that the job is done. The problem surfaces later — in a meeting, during an audit, or when a foreign client queries a term that does not match the original contract. Quality verification exists precisely to catch those errors before the document leaves the building.

What translation quality verification actually involves

Verifying quality is not the same as rereading a text in search of awkward sentences. It is a systematic process that examines specific dimensions: terminological accuracy (do technical terms correspond to the source?), consistency (is the same concept always translated the same way?), completeness (has any segment been omitted?), and register (is the tone formal where it needs to be formal?).

In professional contexts, this verification follows established standards. ISO 17100:2015 defines the requirements for translation services with legal or contractual consequences — including mandatory review by a second qualified translator. ISO 18587 addresses machine translation post-editing specifically, setting out criteria for when and how human intervention should occur.

Why free tools do not verify anything

ChatGPT, Google Translate, and DeepL are useful for understanding the general content of a text. They were not designed to guarantee that a translation is correct — and they do not do so. They produce text; they do not verify what they have produced.

The practical result is a process many teams recognise: someone pastes the text into a free tool, a colleague with partial knowledge of the target language glances over it, and the document moves on. Nobody puts their name to that translation. If an error emerges, there is no record of what was done or who approved it.

Free translation tools also do not preserve the formatting of the original document. A table in a presentation, a header in a report, the structure of a contract in PDF — all of it gets lost or scrambled. The team ends up spending more time reformatting than checking content.

When verification is mandatory and when it is prudent

There are situations where verification is not optional. Documents with legal consequences — contracts, certificates, regulatory submissions, court documents — require certified translation under ISO 17100, with review by a qualified translator and formal accountability for the result.

In other situations, verification is not legally required but is professionally sound:

• Proposals to international clients: an error in register or terminology can damage how the company is perceived.
• Investor materials: inaccuracies in financial documents carry reputational consequences.
• Technical manuals and product sheets: inconsistent terminology creates support problems and may generate complaints.
• Regulatory communications: even where certification is not required, accuracy is expected.

For routine documents — internal reports, operational presentations, day-to-day correspondence — automatic verification is sufficient to ensure a level of quality appropriate for the intended use.

How Vertio builds verification into every translation

Vertio was built with quality verification integrated from the outset, not as an optional step. On the Normal tier (€9/1,000 words), the proprietary engine runs automatic verification checks before delivering the document — and includes a QE report that quantifies translation quality. The Verificada tier (€49/1,000 words) adds review by a human professional for documents that require oversight. The ISO 17100 tier (€89/1,000 words) covers documents with legal consequences, with M21Global certified translators and full compliance with the standard. In every case, the formatting of the original file is preserved — the document comes back ready to use, in .docx, .pptx, .pdf, .txt, .json, and .md formats.

Frequently Asked Questions

What is a QE report in translation?

A QE (Quality Estimation) report quantifies the quality of a translation automatically, identifying lower-confidence segments and terminological consistency metrics. It gives the user a clear indication of reliability before the document is put to use.

What is the difference between automatic and human review in translation?

Automatic review analyses the translation systematically according to programmed criteria — accuracy, consistency, completeness — and is carried out by the translation engine before delivery. Human review involves a qualified professional who reads and corrects the text using linguistic and contextual judgement.

When is ISO 17100 certification required for a translation?

ISO 17100 certification is required for documents with formal legal or contractual consequences — legal contracts, certificates, regulatory submissions, and court documents. In commercial contexts without an explicit legal requirement, the standard may be adopted as an additional quality assurance measure.

Do free tools like Google Translate carry out quality verification?

No. Tools such as Google Translate or ChatGPT translate text but do not perform any verification process on the result. The user receives no indication of translation reliability and no guarantee of terminological consistency.

Does quality verification also cover document formatting?

That depends on the service. Linguistic verification covers textual content — accuracy, consistency, and register. Formatting preservation is a separate function that ensures tables, styles, and the structure of the original document are maintained in the translated file.