Translation quality is one of those concepts everyone mentions but few people actually define. At its core, delivering a quality translation means meeting the client's exact requirements, in terms of language, tone, format, terminology, and intended audience. That sounds straightforward, but it involves two distinct processes that translation agencies use every day: Quality Assurance (QA) and Quality Control (QC).
Understanding the difference between them is not just academic. It is the foundation of how professional translation services operate, and it directly affects the final result the client receives.
What Quality Assurance in Translation Really Means
Quality Assurance in translation is a proactive, process-oriented system. It is put in place before a single word is translated. QA establishes the guidelines, workflows, tools, and documentation that frame the entire project. It defines who does what, by when, using which resources, and according to which standards.
According to ISO 17100:2015, the primary international standard for translation service providers, QA encompasses the complete set of planned and systematic activities implemented to ensure that translation services fulfil quality requirements. In practice, this means evaluating the source text for clarity and completeness, setting up style guides and glossaries, assigning the right linguists for the subject matter, and confirming deadlines before work begins.
In 2025 and into 2026, QA has expanded to include a new layer, managing translation quality assurance tools and AI-assisted workflows. Platforms like Xbench, Verifika, and proprietary CAT-integrated checkers are now standard components of QA setups in professional agencies. A 2025 survey by Nimdzi Insights found that over 68% of mid-to-large translation companies had integrated at least one automated QA tool into their standard project pipeline, up from 51% in 2022.
What Quality Control in Translation Really Means
Quality Control is reactive and product-oriented. It happens after a defined stage of the work is complete, typically after the initial translation draft, after editing, or at final delivery. QC activities include proofreading, error checking, formatting verification, and ensuring that all client-specific requirements set during QA are actually reflected in the finished document.
Proofreading is one of the most critical QC steps, it is the final layer of review that catches residual errors in terminology, grammar, and consistency before the document reaches the client. QC asks practical questions, Does the translated document match the required layout? Have all segments been translated? Are numerical values, dates, and product names accurate? Does the terminology align with the approved glossary? Is the tone consistent throughout?
These are not minor details. A mistranslated dosage in a pharmaceutical document, a wrong figure in a financial report, or an incorrect product name in a technical manual are exactly the types of errors that QC is designed to catch before delivery.
QA and QC Are Not the Same and Cannot Work Without Each Other
A common misconception is that QA and QC are interchangeable terms for the same process. They are not. QA is systemic and preventive, QC is specific and corrective. QA reduces the likelihood of errors by building a sound process, QC catches the errors that still occur despite that process.
The most common real-world failure in translation quality is skipping or shortcutting QA. When a client sends materials with no briefing, no glossary, and no style guide, and expects a finished translation in 24 hours, what the agency performs is QC without QA. The result is that corrections happen at the end, are more expensive, and are more likely to require another round of review.
A 2024 Common Sense Advisory study reported that rework costs due to inadequate upfront quality planning account for 23% of total project costs in translation engagements.
The Role of AI in Translation Quality Assurance Today
One of the most significant shifts in recent years is the integration of AI-generated translation into professional workflows. Machine Translation Post-Editing (MTPE) is now a standard service offering, and it has added new dimensions to both QA and QC processes.
Quality control in AI-assisted translation requires linguists to assess not just accuracy but also the presence of machine-generated artifacts, hallucinations, over-literal phrasing, inconsistent register, and false fluency. A sentence can read perfectly and still be wrong. This has pushed agencies to invest more in QA-level briefings that specify how MTPE output should be treated and what editing standards apply.
The European Language Industry Survey 2025 noted that 74% of respondents had updated their QA and QC protocols specifically to address AI-generated content, compared to 38% in 2023.
Who Is Responsible for QA and QC?
There is no single answer to this, and the honest response is that it depends on the agency's size, specialisation, and project type. In smaller agencies, the same linguist may perform both translation and QC review. In larger organisations, dedicated quality managers oversee QA protocol design, while separate editors and proofreaders handle QC. For documents where accuracy has legal or regulatory implications, contracts, medical records, immigration files, certified translation adds a further layer of accountability, as the translator formally attests to the completeness and accuracy of the translation.
What matters most is that QA is treated as a planning function, not an afterthought. Waiting until the end of a project to define quality requirements is the translation equivalent of building a house without an architect's plan and then hiring an inspector to figure out what went wrong.
So, translation quality assurance and quality control are complementary systems, not competing ones. QA sets the conditions for success before work begins. QC confirms that success was achieved once the work is done. Agencies that invest in both consistently deliver better results, generate fewer revisions, and build stronger client relationships.
If you are evaluating a translation provider, ask them directly, what does your QA process look like before the project starts? The answer will tell you everything you need to know.
Frequently Asked Questions
What is the difference between QA and QC in translation? Quality Assurance (QA) is a proactive, process-oriented system set up before translation begins. It defines workflows, standards, glossaries, and guidelines for the entire project. Quality Control (QC) is a reactive, product-oriented process applied after a stage of work is complete, focused on catching errors before delivery. QA prevents problems from occurring, QC detects the ones that still slip through.
How is translation quality measured? Translation quality is typically assessed against four criteria, accuracy (does the meaning match the source?), fluency (does it read naturally in the target language?), terminology consistency (are specialised terms used correctly and uniformly?), and compliance with client-specific requirements such as style guides and glossaries. Professional agencies use frameworks like the MQM (Multidimensional Quality Metrics) model and tools like Xbench or Verifika to apply these criteria systematically.
How does AI affect translation quality control? AI-generated translations can produce fluent text that is factually or contextually incorrect, a problem known as false fluency. A sentence can read perfectly and still be wrong. QC in AI-assisted translation requires post-editing specialists who can identify machine-generated errors including hallucinations, inconsistent register, and over-literal phrasing that standard proofreading would not catch. This is why agencies working with MTPE workflows need updated QA protocols that specifically address AI output.