Machine translation for a global pharma company

Our client is a medium-sized global pharmaceutical company, who has been working with us for many years.

Towards the middle of 2024, our client got in touch with somewhat of a conundrum. They had a large potential project on their hands – website content and patient materials that they wanted to localize for a number of quite different markets – some European audiences and some Indian.

THE CONUNDRUM

The problem was that there was a lot of content and not much time to get it translated. If they didn’t manage to translate within a certain timeline, these markets would have to be excluded from their trial.

Added to this were budgetary concerns resulting from some unforeseen expenses earlier in the process. As this was a long-term client, we had spoken to them in the past about the opportunities presented by machine translation and AI, and they were curious whether we thought these solutions could help them get this tricky project over the line.

However, this would be their first time working with AI, so they were understandably nervous about not meeting their usual quality standards, and it was clear they would need reassurance and detailed guidance.

To summarize, the challenges of this particular project were:

• Large-scale project

• Tight timeline

• Budgetary issues

• Variety of content types and languages

• Nervous client hadn’t worked with MT before

WHAT WE DID

We started as we would with any MT project; analyzing all content to assess suitability for machine translation. Through an automated process, we tested different machines against different languages and content types to ascertain the best fit for each, and looped in our experts to make the final calls.

As a result of this first phase, we were able to determine which content types and languages could be processed via machine translation, and which would need to go through our standard, human-only process.

As a general rule, our clients can be as involved or uninvolved in this process as they wish. In this case, given their concern over working with MT for the first time, our client wanted to be in the detail with us. And so, we guided them through the process as we went, highlighting and explaining key learnings and decisions at each stage.

This included why, for instance, policy documentation could be processed via MT for most European languages, while subtitling could not. As well as why all the content included in this project would need to go through our standard human process for the relevant Indian languages.

It also included a full rundown of what each workflow stream entailed, with the reassurance that our linguists and experts were in full control of the end product in all cases.

THE RESULT

All content across all workflow streams was completed and delivered on time and within budget.

This allowed our client to include all audiences in their clinical trial. Meaning the efficacy and side effects of the relevant drug will be tested across a much more diverse population.

On our side, we were able to hone a large-scale MT process for a valued partner, and educate and reassure them along the way. And lessons were fed back into our general MT processes, which has informed changes and evolutions to the existing workflows.

And a key indicator that the project was a success? We have gone on to do many more MT and AI-based projects for this client since.

“I really appreciate you looking for ways to get this project over the line, and for your patience in explaining the processes to me. You guys make translations so much less stressful than my experiences before. Thank you!” – Key Account Manager, Global Pharmaceuticals Company

Are you thinking about using MT for the first time? Or wondering if implementing AI might help you get go-aheads you wouldn’t get otherwise?

Get in touch with us for advice, scoping, and step-by-step guidance. We’d be delighted to help.