With all that's being said about AI tools, I firmly believe that if used for the right purpose, they can enhance the productivity of translators. One of these uses is simplifying your marketing efforts by brainstorming ideas and, in the case of this article, creating a whole list of agencies with names and links to their website, so that we can apply more efficiently.
Marketing to agencies is indeed the usual route I pursue when finding new clients online. Although searching on Google or Yandex (!) can return not only single agencies' websites, depending on the keywords you use, but also whole lists with 50 entries, or even pages or pages of agencies to apply to.
Another thing I firmly believe is in the law of large numbers. Simply put, the more people you contact and agencies you apply to, the quicker you'll find the right offer matching your rates and ideal job environment.
ChatGPT is the main AI chat/assistant tool I use for my queries. I used to prefer Gemini simply because of Internet access. But recently, I'm found ChatGPT to be simply better, with Gemini still not remembering details from conversation, fewer customization options, fewer chat handling commands, and the 'free' use of the advanced ChatGPT 4.0 (with Internet access!) multiple times a day, within the token threshold.
One simple, efficient way I use ChatGPT is by asking the following prompt: 'give me a list of companies to cold call within my fields [list of fields I work on]'.
Note that before that, I already fed ChatGPT the way it should answer and most importantly, the information it should know about me when tailoring answers to my requests. To do that, click on your avatar on the chat page, and go to 'Customize ChatGPT'. More information here.
ChatGPT will start giving a list of agencies, actually 5 per field. But how can we make that more efficient? Perhaps by formatting everything in a nice table and including far more data, such as company seat, estimated salary, number of employees, link to website, and most importantly - far more entries, let's say 50 instead of 5?
We can just feed ChatGPT a query like 'now put everything into a table of 200 entries, 50 for each field as per before, and also add the following columns: company seat, estimated salary, number of employees, link to the website'.
Spoiler: the result is limited but still quite astonishing: it generated 80 entries (instead of 200), but they were all credible results. Of course, to avoid hallucinations we should do our research on each, which can go as simple as clicking on each link and see if the website offered exists, and if it matches the info provided by ChatGPT and based on our common sense.
Note: for this example, the agency names and URLs were redacted. Always double-check your sources!