Investing in Language Technology

Unless your organisation has localization leader, to understand the why, when, how and what of acquiring language technology (lang tech), you are going to need help to navigate the process.

What is covered in lang tech?

When we are discussing lang tech, we are simplifying it as connected management and production tools to assist the automation of localizing content. This can include solutions such as a Translation Management Solutions (TMS), Machine Translations tools and an ecosystem of API connectivity.

When considering a TMS, the starting point should be why.

Are you trying to automate your processes?

Do we need business analytics?

How are you tracking spend?

Most companies start the journey by going to their localisation provider to provide a portal (Managed TMS), which can help address some of these challenges. However it won’t enable you to truly manage your content, and you will often face a lot of resistance in change management.

As buyers mature, they tend to move away from a manged TMS to look at agnostic solutions where they can independently manage their providers, control their linguistic assets (translation memories and terminology) and connect with their own technology ecosystems.

In our experience, one of the main considerations is whether you can justify an ROI across all localization spend (translation, technology, interpreting, transcreation, video localization etc), which can generally be justified once a company is managing spend in the hundreds of thousands. When moving to an agnostic TMS it is important to consider that whilst you will have an increased amount of control in the business, it will mean that you will need to dedicate more project management resources to the technology.

We see significant underutilisation of machine translation (MT) and MT post editing in the regulated sector. The only logic we have found is the conservative nature of some firms against regulatory monitoring, security concerns and the perceived quality of MT output.

The subject has been at the forefront of recent discussion with the emergence of Chat GPT and trailing content for automation. Whilst security will always be a concern for regulated customers there are now many solutions available to automate certain content identified during a content audit.

Whether internalising or externalising machine translation, the clear evolution is for certain types of content (identified in your content audit) to have machine translation post edited by humans. This involves human reviewing and improving the MT output rather than raw MT. The benefit to the business includes reduced cost and quicker delivery timelines, with ideal content for this type of work being regulatory materials such as SFDR, KIID, monthly commentaries, financial statements etc.

We are seeing an interesting evolution of source content being draft authored by tools such as Chat GPT with a human review, and the same process should be considered when replicating your language outputs.

Whilst language tech is very important to help clients mature and control their localization spend, real efficiencies, cost savings and risk reduction will only be achieved once you have integrated technology (API) into your existing tech stack. Connected content will minimise the amount of content you need to localise and limit any requirements of file engineering or file transfers. 

Focussing only on outsourcing translations and the price per word will only start the conversation of localization maturity, and building your ecosystem of technology and services providers needs to be integrated into your content and localization strategy.

To learn more, please connect with us..