translation
Our Approach
AI translation keeps getting more accurate, and that progress is steadily reaching patent translation too. In response, ip-pro decides project by project how much to automate and where people take over. Here we set out that approach and how it works in practice.
Choosing the translation method: two modes
We decide the translation method for each project, together with the client, based on its purpose, intended use, and tolerance for risk. ip-pro works mainly in the following two modes.
Human Translation
An experienced patent translator works directly from the source text. This suits projects that turn on claim interpretation, projects anticipating litigation or enforcement, and regulatory work where the tolerance for mistranslation must stay extremely low.
Even in human translation, ip-pro draws on CAT tools and the translation memory it has built up over time to keep terminology consistent for each client and aligned with past projects. Human judgment leads; the tools support it.
Machine Translation Plus Post-Editing (MTPE)
In this mode, ip-pro's translators post-edit the machine-translation or AI output the client provides. As a rule, we work with the client's own translation engine or AI output, in line with their security and operational policies.
Post-editing is done by translators experienced in patent translation and familiar with the relevant technical field, who review and correct the text on the following points.
- Technical accuracy of meaning
- Impact on claim interpretation
- Terminological consistency and alignment with existing translations
- Grammar and style (whether it holds up as a filing document)
This balances cost and quality.
Looking ahead
Looking further ahead, and in response to client requests, ip-pro is also considering building its own machine translation and AI translation environment to offer more flexible solutions. We aim to meet needs such as the following:
- Support that covers both translation and the management of AI translation
- Batch processing in a secure environment
- Advanced integration with translation memory and terminology assets
While preserving the strengths of human translation, we will advance our use of technology step by step.
How we position AI and machine translation
ip-pro does not treat AI or machine translation as a replacement for the translator's judgment. We use them as supporting tools — for studying terms and expressions and for working more efficiently. Whether AI or machine translation enters the workflow depends on the client's request and the nature of the project; ip-pro never introduces them at its own discretion.
Using AI in a supporting role
- We use enterprise generative AI via API, where data is not used for training
- We apply it to studying terms and expressions and to organizing reference material
- We build and share AI assistants aimed at improving the efficiency of our work
- We draw on translation memory and CAT tools to assist the translation using past assets
AI is not what does the translating; it is a tool that supports human judgment.
Handling machine and AI translation
- We can use machine translation engines that do not reuse input data
- We handle post-editing of client-provided machine translation and AI translation output
We provide dependable support for client-led use.
Translation and post-editing by experienced IP translators
Both translation and post-editing are done by translators experienced in patent translation, backed by the team below.
- Each translator covers a broad range of technical fields, spanning several of the electrical, mechanical, chemical, and biotechnology areas
- Translators and project managers hold regular study sessions and meetings to share knowledge and information
ip-pro's post-editing does not stop at fixing grammar or smoothing the wording. Beyond hands-on patent-translation experience, we expect the assigned translators to understand the technical content and claim structure and to judge how the wording affects the interpretation of rights.
Project and delivery management
In patent translation, quality and turnaround depend not only on an individual translator's ability but largely on teamwork and the systems behind it. ip-pro is built so that no single person is a point of failure, while everyone involved takes ownership and works with care — so clients can entrust their projects to us with confidence.
- A project manager (PM) leads each project, working closely with sales, translators, and checkers to oversee the whole and keep it running smoothly
- As the client's point of contact, the PM keeps them informed of progress so they can proceed with confidence
- In line with ISO 17100 certification, we assign PMs who meet the required skill criteria and keep developing them to sustain quality
- We standardize project and delivery management so the work stays consistent no matter who is in charge
- Drawing on what we have learned from experience, we build safeguards into our operations to head off human error before it happens
Underpinning all of this is our translation management system (TMS). Through it, we gain a clear, centralized view of the workflow, enforce strict security through access permissions, check terms and notation for consistency, and align with past projects via translation memory. Pairing human judgment with the system's management and verification keeps our projects running steadily. How we handle data, the international standards involved, and the responsibility ip-pro bears are set out in Our Stance & Responsibility.
Built per project, not uniformly
Rather than applying one method across the board, we choose it to fit each project's purpose and importance. We select either human translation or machine translation plus post-editing (MTPE) to suit the intended use and the client's wishes. ip-pro explains what each involves and then settles on the approach together with the client, following the client's lead.
Details by audience: