ISO 20252:2026 and Translation Management: What Market Research Companies and Language Service Providers Need to Know

The publication of ISO 20252:2026 in September 2026 marks an important evolution in how the market research industry approaches translation of data collection instruments and other project-related documents. Whilst the previous edition of the standard (2019) focused heavily on prescriptive human-led translation workflows, the revised standard adopts a more flexible, technology-neutral approach that reflects the growing use of machine translation, AI tools and automated quality assurance processes.

For both market research companies and language service providers (LSPs), the changes create new opportunities, but they also introduce additional governance responsibilities. 

A Shift from Prescriptive Processes to Demonstrable Outcomes

Under ISO 20252:2019, translation requirements were largely centred on who performed the work and how the review process was structured. The standard required translation to be undertaken by individuals with native-speaker competence or equivalent ability, supported by an independent review conducted by someone other than the original translator.

This approach reflected traditional translation industry practices and provided a clear, process-driven route to compliance.

ISO 20252:2026 takes a different approach. Rather than prescribing a specific workflow, the new version focuses on achieving reliable and accurate outcomes. The standard now recognises that translations may be produced by human translators, artificial intelligence systems, machine translation platforms, or a combination of these approaches.

The question is not “Who translated this document?” but “Can the organisation demonstrate that the translation is fit for purpose and that the process used is reliable?“

AI Translation Receives Formal Recognition

Perhaps the most significant change in ISO 20252:2026 is the explicit recognition of AI and related technologies within the translation process.

Many research organisations and LSPs have already adopted neural machine translation and large language model-based tools to improve efficiency, reduce costs and accelerate multilingual projects. Until now, standards have often lagged behind industry practice.

The 2026 revision acknowledges the reality of modern translation workflows by formally permitting AI-assisted translation. However, this should not be interpreted as a relaxation of quality expectations. 

The standard now requires organisations to demonstrate that any translation solution, whether human or technology-based, is competent and suitable for its intended purpose. In practical terms, organisations must be able to justify why a particular translation approach is appropriate and show evidence that it delivers acceptable results.

Greater Emphasis on Risk and Context

Another important change is the introduction of a risk-based philosophy.

Market research documents are not all equal. A customer satisfaction survey may present a relatively low translation risk, whereas a healthcare questionnaire, regulatory research project or politically sensitive study may require significantly more rigorous controls.

The revised standard requires organisations to consider factors such as:

  • the nature of the source material
  • the sensitivity of the content and
  • the intended use of the translated document.

This aligns translation management with the broader direction of modern ISO standards, where organisations are expected to apply controls proportionate to risk rather than simply follow a predefined process.

For research organisations, this means translation decisions should become part of project planning and risk assessment activities rather than being viewed as a standalone linguistic exercise.

Verification Requirements Become More Flexible

One of the most noticeable departures from the 2019 edition is the removal of the mandatory independent reviewer requirement.

Previously, organisations were expected to have a second person review the translation. Whilst this remains a recognised good practice and may still be appropriate for high-risk projects, the 2026 standard no longer mandates that exact approach.

Instead, organisations must determine the appropriate level of verification and ensure that the translated content accurately preserves the meaning of the original source material.

This gives organisations greater flexibility in designing translation workflows but also places greater responsibility on them to justify their approach.

For example, a low-risk survey translated using a validated AI platform may require a different verification approach from a multilingual medical research questionnaire. Both could be compliant if the organisation can demonstrate that the chosen controls are appropriate.

The Introduction of Translation Governance

A particularly interesting development is the introduction of explicit requirements around translation competence and system governance.

Where AI or automated translation technology is used, organisations must be able to demonstrate that the solution is robust, suitable and capable of delivering the required quality level.

This moves the discussion beyond translation itself and into the wider field of technology governance.

Organisations will need evidence demonstrating that translation tools have been selected appropriately, evaluated effectively and monitored over time. This is consistent with broader regulatory and standards developments surrounding artificial intelligence and responsible technology use.

For many organisations, this may involve supplier assessments, benchmarking exercises, quality monitoring activities and documented validation processes.

Defining and Measuring Accuracy

The 2026 edition also introduces a concept that was largely absent from the previous version: defined translation accuracy levels.

Rather than relying solely on prescribed process controls, organisations are expected to establish quality expectations and verify that these are being achieved.

This represents a move towards evidence-based quality management. Translation quality can no longer be assumed simply because a particular process was followed. Organisations should be capable of demonstrating that translated content achieves the required level of accuracy and meaning preservation.

For language service providers, this creates an opportunity to offer measurable quality assurance frameworks and performance metrics as part of their service offering.

Where Does ISO 17100 Fit?

For language service providers looking to demonstrate compliance with the new ISO 20252 translation requirements, ISO 17100 remains highly relevant.

ISO 17100 is the internationally recognised standard for translation services and provides detailed requirements relating to translator competence, revision, resource management and quality assurance.

Although certification to ISO 17100 is not explicitly required by ISO 20252, organisations that maintain ISO 17100 certification already operate many of the controls expected by the revised research standard. For example, ISO 17100 requires the use of competent translators, documented project management processes and independent revision activities.

As a result, certification to ISO 17100 can provide powerful supporting evidence that an organisation has a structured and controlled translation management system.

ISO 17100 can also be paired with ISO 18587, the international standard for full post editing of machine translation output.  This does some way to addressing output of machine translation and AI tools, although both these standards are under review.

For research companies selecting translation partners, working with an ISO 17100-certified provider may also offer additional confidence that translation activities are being managed through a recognised quality framework.

Translation Only part of the Solution

While ISO 20252 addresses translation specifically, it does not recognise the important role of localising content to cultural context and sensitivities for the most accurate research data from participants.

This is something that, currently, is difficult for machine-translation and AI tools to achieve.  Depending on the type of content being using, market researcher firms may still benefit from a human review or advice when fielding studies in certain territories.

How Assent Risk Management Can Help

At Assent Risk Management, we support market research organisations and language service providers in understanding and implementing management system standards. Whether you are preparing for ISO 20252 certification, transitioning from the 2019 edition, evaluating AI-enabled translation workflows or seeking alignment with ISO 17100, our consultants can help you develop practical, audit-ready systems that meet the intent of the standard while supporting operational efficiency.

The 2026 revision reflects the reality of today’s multilingual research environment. Organisations that successfully combine robust governance with modern translation technologies will be best placed to demonstrate compliance, maintain client confidence and deliver high-quality research outcomes across global markets.

TopicISO 20252:2019ISO 20252:2026Change
Who may perform translations?Translation performed by individuals with mother-tongue competence or equivalent skills in source and target languages.Translation may be undertaken by competent individuals or AI/technology-based solutions.Expanded to explicitly include AI and automated translation technologies.
Translator qualificationsAt least one translator must have experience with the type of document being translated.Service provider must ensure competence of individuals or suitability and competence of translation systems.Moves from translator qualifications to outcome and competency assurance.
Machine Translation (MT)MT permitted but subject to human checking and revision requirements.AI and related technologies formally recognized as legitimate translation methods.AI/MT becomes explicitly accepted rather than an exception.
Independent review requirementTranslation must be checked and revised by someone other than the translator.Verification process determined by the organisation based on need and risk.Mandatory second-person review removed.
Verification methodPrescriptive checking and revision process.Flexible verification approach based on risk and intended use.Shift from prescribed process to risk-based validation.
Preservation of meaningImplicit through review requirements.Explicit requirement to ensure translated content accurately preserves original meaning.Strengthened and made explicit.
Risk assessmentNo specific translation risk assessment requirement.Translation and verification approach must consider sensitivity, purpose and intended use of translated material.New requirement.
Research instrument translationGeneral translation requirements applied.Special attention expected for questionnaires, discussion guides and materials affecting research outcomes.Greater emphasis on research validity.
Governance of translation processPolicy required where translator competence could not otherwise be verified.Systematic process required to demonstrate competence of translators or robustness and suitability of AI solutions.Much stronger governance requirements.
Translation quality targetsNo explicit accuracy targets required.Defined translation accuracy level required, with a process to verify achievement.Entirely new requirement.
Records of review activitiesRecords of checking and revision required.No explicit requirement for revision records within this clause.Specific record-keeping requirement removed.
Client reviewClient given opportunity to review translations.Client may review translations and may request information on translation and verification methods used.Greater transparency.
Overall approachHuman-centric process control.Technology-neutral quality and governance control.Major philosophical shift from process-based compliance to outcome-based assurance.
Robert Clements
Robert Clements
Articles: 373