TL;RD

  • Microsoft Power Platform may be an interesting way to replace many features found in Translation Business Management Systems (TBMS).
  • A TBMS may be too expensive for small and medium language services providers, and introduce many limitations.
  • You can increase your business productivity and reduce overhead costs by an important factor by encouraging low-code/no-code training in your organization.
  • The integration of AI in low-code/no-code platforms may completely change the way we interact with systems supporting localization operations.

TBMSs are great, but there’s a problem.

If you work or have worked for translation companies, you have realized how fundamental tools such as XTRF, Plunet, LSP Expert, Protemos, and others are to keep track, monitor, and report all money, project, and vendor operations.

These tools replace the traditional Excel or Google Sheets spreadsheet with giant relational databases featuring user-friendly filtered views and forms.

For instance, you can opt to only see those columns associated with vendor information for one view and project information for another view. In both cases, you will have a form to introduce a new row of data.

Project managers, accountants, clients, and vendors feed these databases with new data.

The problem is their cost, especially if you run a team of several project managers.

For instance, for a team of seven, your average yearly operational costs could be around US$18,000–US$20,000.

Add to this the cost of a Translation Management System (starting at US$15,000 for the same team).

Yes, US$35,000/year is a lot. And there’s a problem.

Although most of these platforms claim that they reduce the friction of using a spreadsheet to run multiple operations, they introduce a new type of manual endeavor through a different interface and establish limitations that didn’t exist in the overly loose yet impractical spreadsheet.

I know what you are thinking: “Some of them have vendor and client portals to partially automate part of the data entry work (e.g., payables, receivables, invoices, due times, language pairs).”

But what if the clients don’t want to use your portal but email?

What if you want to customize those portals because their logic doesn’t serve your business workflows?

That’s when I asked myself – “Why pay that much if you can replicate the way a TBMS operates using low-code apps?”

Low-code and no-code applications as an alternative to TBMSs

In 2021, I built a series of admin automations for The Translation Team using Microsoft Power Automate.

The goal was to stop moving or copying a specific set of files from here to there manually. It’s nothing fancy, but hell, we saved a few clicks.

That was one of my first experiences with the concept of workflows, automations, and low-code/no-code applications.

Over two years, their library of automations has grown significantly – custom Teams notification when something happens somewhere, automatic emails based on specific criteria, that kind of stuff.

But there was one project that demonstrated how impactful these low-code platforms could be.

In November 2022, The Translation Team welcomed Rosalía Mognol, former vendor and project manager for Welocalize Life Sciences, to continue the expansion of the project management team.

She had this idea of centralizing and streamlining recruitment and onboarding tasks to quickly support project managers when dealing with multilingual projects or projects where subject matter expertise was key to success.

The challenges were clear once we crystallized the idea into a project: there were numerous manual tasks, such as document handling, reminders, confirmations, notifications, signing forms, and data entry in different systems.

Long story short. We made it. We built this recruitment/onboarding process by gluing different platforms with multiple automations, overcoming most of the forecasted doomy manual tasks.

It still has a long way to go, but I haven’t seen anything like this in a language services provider.

If you could picture this process in a diagram, you would see a big table (database) in the center, with multiple arrows coming in and out, establishing relations.

Just like the ones I mentioned at the start of the post.

Isn’t that the principle of Plunet, Protemos, XTRF, or LSP Expert?

If we succeeded in building such a process, what would it take to create another one for projects and finances?

Use your imagination.

Raise citizen developers in your LSP

If there is a term Microsoft has coined that I love, it is Citizen Developer.

It means that you can code something that will have a big impact in your org without having an MA in Computer Science.

Imagine the business challenges you could solve if key members from each department in your LSP or localization department (buyer side) knew how to build things like this recruitment/onboarding system.

Yes, mindblowing. And it would give everyone more agency and a sense of belonging.

The good news is that it doesn’t take too long to achieve this. Microsoft Learn has a plethora of learning paths with various levels of difficulty.

They can start from scratch and evolve into pro low-code citizen developers.

And who knows, it may be love at first sight once they see what they can create for their peers and themselves.

And with AI basically plugged everywhere, it can make this process much faster and enjoyable.

The impact of Large Language Models in low-code apps

Since we built our recruitment and onboarding system in Microsoft Power Automate, it means that we can further improve and expand its features by integrating pre-built AI models into it.

As you may know, Microsoft acquired part of OpenAI in a multibillion-dollar deal last year, so we now have full access to any GPT flavor.

One of the first ideas that I have is resume summarization.

For instance, use Azure Document Intelligence to extract specific entities (i.e., work experience) from resumes in any format (you need to train a small model for this) and have GPT structure them so we can easily—and automatically—import them into our databases.

This would accelerate the review process during onboarding.

You could do the same to calculate gross margins, profits, COGS, and many other money things.

The smooth business operations you’ve always dreamt about

My hope is that all the above has gotten you excited and thinking about how you can use these tools to improve your business processes and operations.

The example I gave you using TBMSs is just one of many other use cases, but since we are in the middle of an economic crisis globally, it makes sense to cut costs where it makes sense.

Not in professional services, meaning paying less to translators, voice artists, subtitlers, and copywriters, but in those dumb admin tasks that steal us so many hours (and dollars).

Конец.