7 mistakes that companies make when translating medical or technical content
Localisation of medical or technical content is not a formatting task; it is a risk-management process. When companies treat it as a late, low-cost or automated step, they increase the chances of errors that can affect approvals, safety, credibility and business performance in regulated markets. In sectors such as pharma, medical devices, engineering, energy, fintech and defence, translated content is often read by people who must act on it: clinicians, patients, regulators, technicians, legal teams and buyers. A mistranslated instruction, inconsistent term or unvalidated AI output can create consequences far beyond awkward wording. The seven mistakes below are common because they look efficient at first. In practice, they often create rework, delays and exposure that could have been avoided with the right process from the beginning. 1. Why is treating translation as an afterthought so risky? The first mistake is leaving Localisation planning until the end of the content workflow. By then, the source text may already be approved, designed, locked in complex formats or distributed across several teams and platforms. This creates avoidable pressure. Translators and reviewers are asked to deliver faster, with less context and fewer opportunities to ask questions. In regulated industries, that is exactly when small inconsistencies can become serious quality issues. Translation should be planned from day one, especially when content will support product approvals, patient information, technical operations or contractual commitments. Early involvement allows terminology, reference materials, workflows and review responsibilities to be defined before deadlines become critical. A better approach starts with questions such as: Which markets, languages and regulatory contexts are involved? Who will approve the translated content internally? Which terms must remain consistent across documents? Which files, formats and platforms will be used? What is the acceptable review and validation process? When translation is integrated early, teams reduce last-minute decisions and protect both quality and time-to-market. 2. What happens when companies choose providers based on price alone? The second mistake is treating the lowest quote as the safest business decision. In complex medical or technical content, price-only selection often hides the real cost of poor quality. A cheaper supplier may lack specialist linguists, structured quality assurance, terminology controls or experience with regulated documentation. The result can be rework, internal escalation, delayed approvals and loss of confidence among stakeholders. This does not mean that cost is irrelevant. It means cost must be assessed alongside risk, expertise, process and accountability. Translation for a brochure, an internal memo and a regulated product document should not be bought as if they carried the same consequences. Companies should evaluate providers on practical criteria: Proven experience in medical, technical or regulated content Clear quality assurance and review stages Ability to manage terminology and reference materials Secure handling of confidential documentation Responsiveness when questions or deviations arise The right question is not “Which supplier is cheapest?” but “Which supplier reduces the risk of error, delay and non-compliance?” 3. Why does subject-matter expertise matter so much? The third mistake is assuming that fluency in two languages is enough. It is not. Medical and technical translation requires subject-matter expertise, because the translator must understand both the language and the content logic behind it. A generalist may produce text that sounds natural but fails technically. In regulated communication, that is especially dangerous because the error may be subtle: a dose, device component, procedural step, claim, limitation or warning may be expressed in a way that changes its meaning. Medical translation requires control of terminology, patient safety implications, document conventions and the difference between technical precision and plain-language clarity. Technical translation requires understanding of systems, processes, components, operating conditions and the practical consequences of ambiguous instructions. Regulated-sector translation requires awareness that every word may be reviewed, queried or relied upon by a professional, auditor, regulator or end user. Subject-matter expertise also improves collaboration. Specialist translators know when to ask a question, when a term should not be “improved”, and when consistency matters more than stylistic variation. In regulated content, a translation is not successful because it reads well; it is successful because the right reader can act on it safely and correctly. 4. How does inconsistent terminology create compliance problems? The fourth mistake is allowing different terms for the same concept to appear across related documents. Terminology consistency is not a cosmetic preference; it is a core quality requirement. If one document uses one term, another uses a near-synonym and a third uses an internal abbreviation, readers may wonder whether the content refers to the same concept or to different ones. In technical and medical contexts, that uncertainty can affect interpretation, training, reporting and regulatory review. Terminology management helps prevent this. It gives translators and reviewers an approved reference point for critical terms, product names, procedures, warnings, component labels and recurring phrases. Good terminology work typically includes: Approved glossaries for key terms and expressions Clear rules for product names, abbreviations and units Alignment across departments and document types Updates when source terminology changes Reuse of validated terms in future projects The benefit is cumulative. Each approved term strengthens future work, reduces review cycles and protects consistency across markets. 5. Can AI translate medical and technical content safely? The fifth mistake is over-relying on AI without expert human validation. AI can be useful for scalability, drafting and workflow support, but human validation remains critical where accuracy, liability and compliance are involved. AI systems can produce fluent output that looks convincing even when it is wrong. That is a particular risk in medical and technical content, where errors may not be obvious to a non-specialist reviewer. A sentence can sound polished while misrepresenting a procedure, caution, limitation or technical relationship. The question is not whether AI should ever be used. The question is where it fits safely inside a controlled process. In regulated content, machine-assisted workflows need expert review, terminology control, confidentiality safeguards and clear responsibility for final approval. Companies should be cautious when AI is used for: Patient-facing information or instructions Clinical, legal or regulatory documentation Safety warnings and adverse-event language Technical
How poor localisation can delay regulatory approval
Localisation can delay regulatory approval when translated content creates doubt: inconsistent terminology, mismatched labels, unclear instructions or missing updates across language versions. In regulated sectors, authorities review multilingual submissions for clarity, consistency and compliance, not just linguistic equivalence. If one version does not align, approval can slow down while teams correct, justify and resubmit content. Why does localisation matter to regulators? Regulators do not assess only the source file. They review the materials submitted for their market, in their language, against the expectations of their regulatory framework. That means your translated content is not a support asset; it is part of the evidence they use to judge whether a product, service or filing is clear, controlled and compliant. In practice, multilingual consistency is a regulatory signal. If terminology changes from one document to another, or if a warning sounds weaker in one language than in another, the submission can appear less controlled. Even when the underlying product is sound, the documentation may raise questions. This matters across regulated sectors where documentation carries legal, technical or clinical weight. Pharma product information, fintech customer disclosures, legal agreements, defence specifications and energy compliance files all depend on precise language. The risk is not that a sentence sounds awkward; the risk is that it changes interpretation. Where does poor localisation usually go wrong? Poor localisation creates uncertainty where regulators expect certainty. The most common failures are rarely dramatic at first glance. They are small inconsistencies that accumulate across documents, labels, forms, instructions and updates. Typical problems include: Different translations for the same regulated term across files. Dosage, safety or usage instructions that are grammatically correct but ambiguous. Labels that do not match the wording used in the supporting dossier. Updates made in one language but not reflected across all markets. Local requirements overlooked because the process treated translation as a late-stage task. The issue is not only linguistic quality. It is regulatory alignment. A translation can be fluent and still be wrong for approval if it ignores approved terminology, national requirements, product history or the relationship between documents. This is why regulated content needs more than a general language review. It needs controlled terminology, market awareness and documented quality checks. Without that structure, teams may discover the problem only after the submission has already been questioned. How can one inconsistent term delay approval? One inconsistent term can force a reviewer to ask whether two documents refer to the same concept, the same product feature, the same obligation or a different requirement. That question creates friction. Once doubt enters the review process, the burden shifts back to the submitting company to explain and correct. Terminology control is therefore not cosmetic. In regulated environments, terms carry approved meaning. A product component, adverse event, financial risk, contractual obligation or safety instruction may have a preferred wording because that wording has already been validated. If a term is translated one way in the product information, another way in the labelling and a third way in the supporting submission, the reviewer has to interpret the relationship between them. That interpretation should never be left to chance. In regulated markets, localisation fails when it makes reviewers work harder to trust what should already be clear. The same applies to formatting, references and version history. If a change is introduced in one language and not mirrored in another, the issue may look like weak document control. The regulator may then ask for clarification, updated files or evidence that the inconsistency does not affect compliance. What is the real cost of delayed approval? When localisation fails, the consequences extend beyond translation. A delayed approval can postpone a launch, disrupt market entry and put commercial plans under pressure. It can also force internal teams into urgent rework when they should be focused on execution. The operational cost is often felt across several functions at once. Regulatory affairs must answer questions. Legal or compliance teams may need to review revised wording. Product, marketing or medical teams may have to confirm which language reflects the intended meaning. Project managers then have to coordinate corrections under tighter timelines. The commercial cost is just as important. In competitive markets, being late can mean losing momentum, missing a planned launch window or giving competitors more time to establish themselves. A localisation issue may appear small, but its effect can move through the whole approval chain. The deeper problem is that preventable rework consumes expert time. These are usually not generic documents that anyone can fix. They require people who understand the product, the terminology, the local market and the regulatory context. Why is treating translation as the final step risky? Many companies still create content in one language, finalise it internally and then send it for translation near the end of the process. In regulated sectors, that model is fragile. It leaves too little time to solve terminology issues, align local requirements or check consistency across all materials. The better approach is localisation by design. That means planning multilingual content before the last mile, not after it. Teams should know which markets are involved, which documents are connected, which terminology is controlled and which reviewers must validate the final versions. A stronger workflow usually includes: Approved glossaries and style rules before translation begins. Clear ownership for terminology and regulatory decisions. Version control across every language and document type. Early input from subject-matter experts. Quality assurance checkpoints before submission. This does not slow the project down. It reduces avoidable stops later. When localisation is embedded early, questions are resolved before they become regulatory queries. Can AI prevent localisation errors in regulated content? AI can accelerate parts of multilingual production, but it cannot ensure compliance on its own. It can support drafting, comparison, terminology suggestions and workflow efficiency. However, regulated content requires accountability, context and validation. Where does the responsibility sit if an AI-generated phrase changes legal or clinical meaning? That is the question regulated companies cannot ignore. In high-risk content, speed is useful only if the output remains