Language technology
A translation platform, integrations and cloud localisation built for scale.
Get a quoteTranslation technology has a reputation problem, and it earned most of it. A decade of tools promised to remove the human and delivered content nobody could publish. The pendulum then swung to the opposite error: organisations translating the same 40 pages by hand, every year, from scratch, because technology felt risky.
Both are expensive. The useful position is in between, and it's not complicated: technology should handle everything that is repetition, routing and version control, and humans should handle everything that is meaning.
CORE+, and what it does
CORE+ is our translation management platform. It sits between your content and our translators and does the work that no one should be doing in an inbox.
- Translation memory. Every segment we translate for you is stored and matched against future content. Repeated text isn't retranslated, so it isn't re-billed, and it can't come back worded differently.
- Terminology management. A termbase of the language you can't afford to see drift: program names, entitlement descriptions, clinical terms, the word your community uses rather than the word your policy uses.
- API and CMS connectivity. Content moves from your system to translation and back without a human exporting a spreadsheet. Translation happens alongside your release cycle instead of after it.
- Real-time project tracking. What's in progress, what's in revision, what's late and who has it. Visible to you, not on request.
- Secure file handling. ISO 27001 certified, which is the answer government IT security assessments are looking for.
Where AI genuinely helps, and where it doesn't
We use machine translation and AI. We're specific about where.
It works for triage and gist: an inbound email, a complaint, a document someone needs to understand right now to decide what to do with it. It works for high-volume content under human review, where a trained linguist post-edits rather than starting from a blank file. It works for internal material with low consequence and high turnover.
It does not work unsupervised for anything with legal, clinical, financial or reputational weight. Not because the output reads badly, but because it reads well while being wrong, which is the harder failure to catch. A confident mistranslation of a dosage instruction or a tenancy right is worse than an obvious one.
For nib Travel Insurance we built exactly this boundary into the workflow: a custom AI-driven translation solution integrated into claims operations, giving hundreds of agents instant access to medical document translation inside a secure, ISO certified environment, with human translation for everything that leaves the building. How that was built.
Websites and digital, without a second CMS
The most common multilingual website problem isn't translating the site. It's the twelfth time you translate it, after content has changed on 40 pages and nobody can tell which.
We connect to your CMS and treat translation as part of publishing rather than a project that runs beside it. Changed content is detected, sent, translated and returned into the same fields. In-language pages carry the right hreflang, the right canonical tags and the right structured data, so search engines index them as alternates rather than duplicates, which is the technical failure that quietly wastes most multilingual site budgets.
The same applies to apps, e-learning platforms and document generation systems. If it produces content on a schedule, it can be connected.
What this changes on a large program
For Uniting we've run an enterprise translation partnership for more than a decade, across 550 services in New South Wales and the ACT. At that scale the value isn't speed on any single job. It's that a translation delivered in 2016 and a translation delivered this month use the same terms, that nobody re-pays for content already translated, and that a new project manager can see the state of every job without asking. How the partnership runs.
Integration work starts with technology integration, and the human quality assurance underneath it is translation services. One without the other is the mistake.
Common questions
Speak to our teamWhat is CORE+?
CORE+ is LEXIGO's translation management platform. It holds your translation memory and terminology, connects to your systems by API, tracks every job in real time and handles files inside an ISO 27001 certified environment. Clients use it to stop paying twice for the same content and to stop chasing status by email. It supports the human process rather than replacing it.
Is machine translation good enough for public-facing content?
Not on its own. Raw machine output is fluent enough to look publishable while carrying errors a reader can't detect, which is precisely the wrong risk profile for health, legal or government information. Where volume makes full human translation impractical, we use machine translation with human post-editing by a qualified linguist, and we tell you which content took which path. Anything certified, clinical or legal is human translated and revised under ISO 17100.
Can you integrate with our CMS?
Yes. We connect by API to the common content platforms and to custom systems, so content moves to translation and back into the same fields without an export step. Where an integration isn't practical, a translation proxy can serve in-language versions of your site without touching the underlying build, which is usually the fastest route to a multilingual site when development capacity is the constraint.
How is our data secured?
ISO 27001 certification covers the information security management system, including access control, encryption in transit and at rest, translator confidentiality agreements and audited handling procedures. Data residency and retention are set per client, which matters for government work with specific hosting requirements. We're asked these questions on every government tender and we're happy to answer them before one.
Does translation memory actually save money?
Yes, and the saving compounds. The first year builds the asset and looks like ordinary translation cost. From the second year on, every repeated segment is matched rather than retranslated, and content sets with high internal repetition, such as policy documents, product descriptions and annual updates, see the largest reduction. The memory belongs to you, not to us, and you can take it with you.
What about hreflang and multilingual SEO?
We handle it, because it's usually where multilingual sites lose the value they paid for. In-language pages need correct hreflang annotations, self-referencing canonicals and localised metadata, or search engines treat them as duplicates and serve the English version to everyone anyway. We also localise the metadata and structured data rather than translating it literally, since search behaviour in Vietnamese or Arabic doesn't mirror the English query.