Insights

AI is no longer a promise for the future of healthcare; it has already become a powerful ally in how the NHS delivers care today. From speeding up cancer detection to predicting cardiovascular risks before symptoms appear, artificial intelligence has moved out of the research lab and into clinical reality.  But while the conversation often focuses on algorithms, processing power, and clinical accuracy, there is a less flashy (yet equally vital) factor that determines whether these innovations succeed: localisation.  Localisation is about much more than translation. It is the adaptation of language, cultural context, and regulatory requirements to ensure that health data platforms can be trusted and effectively used across diverse NHS settings. Without it, even the most advanced AI technology risks exclusion, misunderstanding, or rejection.  The NHS has become a global leader in AI adoption. In 2019, the NHS AI Lab was launched with a £250 million investment, designed to accelerate the integration of cutting-edge AI tools into frontline care. Since then, more than 80 technologies have received support through the AI in Health and Care Award.  The momentum has only grown. In 2023, NHS England announced a further £21 million in funding to expand the use of AI in cancer, cardiovascular disease, and stroke care, according to the UK Government. These investments reflect a recognition that AI can save time, reduce pressure on clinical staff, and, most importantly, improve patient outcomes.  Yet, as AI-driven platforms integrate into patient apps, clinical dashboards, and NHS electronic systems, a critical barrier emerges: ensuring every patient and clinician can engage with these tools effectively.  The UK is one of the most linguistically diverse countries in Europe. The Office for National Statistics (ONS) 2021 Census revealed that approximately 1 in 10 people in England and Wales do not use English as their primary language. For the NHS, this is not a peripheral issue. It is a central part of equitable care.  NHS Digital guidelines clearly state that digital health tools must be accessible to all patient populations. That means AI platforms must support localisation across:  Without this, the NHS risks reinforcing health inequalities and potentially exposing patients to harm due to miscommunication.  The dangers of neglecting localisation are not theoretical.  A BMJ Open study on Miscommunication in Healthcare (2022) highlighted that patients who do not fully understand healthcare instructions are more likely to experience adverse outcomes and lower treatment adherence. This is not only a patient safety issue, but it also increases pressure on already stretched NHS services.  The Medicines and Healthcare products Regulatory Agency (MHRA), which governs medical devices and health technologies in the UK, requires that documentation and labelling are linguistically and culturally accurate. A lack of compliance here can lead to delays in approval and block NHS adoption entirely.  Moreover, Deloitte’s 2023 digital health report emphasised that localisation is “as critical as technical performance” when it comes to patient engagement. Without cultural adaptation, even the most technically sound AI solution may fail to gain traction.  And let’s not forget the financial implications. Failure to localise early often leads to expensive rework, delayed product launches, and strained relationships with regulators and healthcare providers.  For AI-driven health platforms, localisation should not be seen as an afterthought. Instead, it must be built into the development process from day one.  This approach offers several strategic advantages:  One area where localisation makes a measurable impact is digital consent. Many AI health platforms now integrate eConsent modules, allowing patients to review, understand, and agree to treatments or clinical trial participation digitally.  If this content is not accurately localised, patients may misunderstand risks, rights, or procedures, potentially invalidating consent. This is not only an ethical issue but also a regulatory one. Inadequate localisation in eConsent has already led to trial delays and compliance failures in several EU countries.  For the NHS, which is increasingly using digital consent processes to streamline workflows, accurate localisation is therefore essential.  At Avantpage Life Sciences, we understand that the future of healthcare innovation depends on bridging technical excellence with patient accessibility. We work with AI-driven health data platforms to embed localisation at every stage of development.  Our solutions include:  By partnering with us, AI health platforms can accelerate NHS adoption while avoiding regulatory pitfalls, delays, and unnecessary costs.  AI is not just a tool for the NHS; it is rapidly becoming an operational backbone. From predictive analytics to patient-facing apps, these technologies are poised to define the future of healthcare delivery.  But technical brilliance alone is not enough. Success depends on whether innovations are accessible, understandable, and compliant across diverse populations. Localisation is therefore not a “nice to have”. It is a strategic necessity for every AI health data platform seeking NHS integration.  At Avantpage Life Sciences, we are committed to helping organisations cross this bridge. With our expertise in regulatory compliance and patient-focused localisation, we enable innovators to deliver real impact where it matters most: at the point of care.  If you are developing or scaling an AI-driven health platform, consider how localisation can accelerate your NHS journey. To learn more about our work and explore how we can support your goals, contact the Avantpage Life Sciences team. 

In the year since OpenAI launched ChatGPT in November 2022, we humans have been scrambling trying to figure out just what to make of artificial intelligence (AI). No matter the industry you work in — government, medicine, or even language services — the use of AI as a tool has been a hot topic for all of us this year. In the language services industry, we’ve been dealing with AI and translation for even longer now — since the rise of neural machine translation in the 2010s, AI has been an extremely powerful translation tool. But the highly fluent nature of texts produced through modern machine translation tools raises the question: When is human translation better than AI translation, and vice versa? The answer is complicated — it all depends on the type of content you need translated and the specific requirements you have for that translation. That’s why we’ve devised the following roadmap for you to determine whether you should request AI-powered translation, human translation, or a hybrid model for your language services. Here, we’ll take a look at the questions you should ask yourself before determining which approach to use in your translation needs.
The first question you’ll want to ask yourself is about the type of content that you need translated — AI is especially good at translating more technical, repetitive texts, however it still lacks the human touch necessary for more creative types of content, like pun-heavy marketing materials or nuanced blog posts, just to name a couple of examples.  While a human will generally review AI translations before you get the final product, you may find that the translation process goes more smoothly when you start with a human translator from the get-go for certain types of content. You’ll also want to consider some of the specific requirements of the project when you’re wondering which route to go. Here are just a few factors to look at: If you have a tight deadline for a given project and the content type is suitable for AI translation, AI could be the right way to go. AI tends to be much faster, where human translators need more time. That said, if the content type isn’t ideal for AI translation, you may find that the review process ends up taking longer than desired — in cases where your content needs to be translated quickly by a human, consider requesting a rush translation from a trusted language service provider. Human-produced translations are going to cost you more than an AI translation. Still, it’s important to consider the fact that human translations are typically higher quality, and mistakes resulting from an AI translation could be costly. Consider the type of content you need translated first before determining what your budget constraints allow for.
Quality expectations vary across projects — choosing between AI and human translation may also hinge on factors like the expected accuracy, cultural sensitivity requirements, and the project’s potential impact on organizational outcomes and performance. Generally speaking, human translators will be able to ensure more accurate results, as AI often makes errors in vocabulary and context. Additionally, critical content like legal or medical documents demands the precision that human translators provide. AI, while advanced, often lacks the contextual understanding required for such materials. Language service experts often note that AI translates words, not meaning, while humans translate meaning, not words. This is particularly important to keep in mind with texts that may have culturally sensitive information — certain words and phrases might come across bluntly or insensitively when directly translated into another language, and as such, a human translator is important when considering cultural sensitivity. Human translators excel in capturing cultural nuances, ensuring that the translated content aligns seamlessly with the target audience. For projects where translation quality directly impacts outcomes — think educational materials that will assist a patient in their health decisions, or a child’s individualized educational plan that will determine their educational path — investing in human expertise is generally the more strategic choice.
While AI can produce content that meets your voice requirements, human translators are generally much better at creating content that adheres to your style and voice guidelines. A common complaint about AI translation tools is that they can’t consistently account for stylistic requirements — for example, if your brand’s content strictly follows the Chicago Manual of Style, AI may not be able to adequately follow those style guidelines. While machine translation glossaries and adaptive machine translation can be useful for AI translations, human translators still have an edge here. Humans are more well-equipped to translate jargon-heavy or highly stylized texts, and it’s much easier to get them to review and follow your organization’s unique style and tone guidelines.
Because AI-powered translation tools tend to struggle with accuracy more than human translators do, it’s important to consider the risks associated with your project — that is, what could go wrong, and how can you prevent the likelihood of that happening? For highly sensitive content — medical documents, legal texts, and anything else that’s heavily regulated — human translators provide the necessary expertise to navigate complex terminologies accurately. Inaccurate translations of these kinds of texts can be costly, have legal implications, and could even be life-threatening — that means it’s important to work with a trusted human translator who can provide the most accurate translation possible. If you determine that the overall risk is low and opt for machine translation, you can mitigate risk even further by incorporating thorough review and quality assurance processes. Quality assurance tools can perform objective measures of the quality of a given translation, while human reviewers are an absolute must, as they can fill in the gaps in any AI translation. Additionally, user feedback can play a useful role in mitigating the risk associated with AI or human translation. Consider testing out a translation with a small focus group to hear their thoughts on a given translation — this will give you a sample of what end users will think when they encounter the final product, allowing you to tweak details as necessary.
Before you decide to opt for an AI translation tool over a human translator, ask yourself the following questions: Whether or not you use AI or human translation depends on a wide range of factors — and sometimes, the answer still isn’t clear. In those instances, you may find that a hybrid approach works best. At Avantpage, we recognize the value of both AI and human capabilities. Our approach integrates AI features, human expertise, and hybrid methods to ensure a customized solution that works best for your project’s unique demands. If you’re looking to learn more about whether AI translation, human translation, or a hybrid approach is best for your project, contact us today at [email protected] or submit an enquiry. We’ll help you determine and execute the most effective strategy for your project.