|
By Volkan Güvenç, Founder — Alafranga Language Solutions I started Alafranga in 2002. The first tool I used was a fax machine. Not metaphorically. An actual fax machine, for receiving source files and returning translations. Within a year, we moved to email. Within a few more, to FTP transfers and CAT tools. The technology kept changing. The question was always the same: do we get ahead of this, or do we wait and see? Every time we waited, we lost ground. Every time we moved early, we built something. |
▮The CAT tool decision
When SDL Trados became the industry standard in the mid-2000s, there was resistance across the market. Translators complained it devalued their work — fuzzy match discounts felt unfair, repetition penalties felt punitive. Some agencies refused to adopt it, positioning themselves as "premium human translation" providers who did not use CAT tools.
Most of those agencies are gone.
We adopted CAT tools early, not because we thought they were perfect, but because our clients needed terminology consistency across large document sets and we could not deliver that reliably without them. The technology served a real operational need.
▮The ]Project Open[ decision
In 2007, we implemented ]Project Open[ — an open-source project management platform. At the time, no other Turkish translation company was using it. Most were running on spreadsheets and email threads.
The decision was not about prestige. It was about visibility. When you are managing dozens of projects simultaneously across multiple language pairs, you need to know at any moment what is assigned, what is in review, what is overdue, and what the client has not approved. A spreadsheet cannot tell you that.
Twenty-three projects into that year, we could see patterns we had never seen before. Which document types were causing delays. Which language pairs had the most revision cycles. Which clients were sending unclear briefs. The data was there. We just finally had a way to read it.
▮The MT post-editing decision
In 2018, we introduced machine translation post-editing into our workflows. This was more contentious than the CAT tool decision had been.
Some of our best translators pushed back. They felt MT output was beneath them — that reviewing a machine draft was not real translation work. Some refused to work on MT projects at all.
I understood the concern. But the clients asking for MT post-editing were not looking for a cheaper product. They were looking for a faster one on content that was suitable for it — internal communications, knowledge base articles, product descriptions in low-risk markets. The content did not warrant full human translation. The question was whether we would be the ones to deliver it well, or whether the client would find someone who would.
We built a workflow that made the distinction explicit: which content types were suitable for MT post-editing, which required full human translation, and why. That framework became our SmartEdit workflow. It is still in use today, expanded to include AI drafting alongside MT.