Technical Articles
Translating PDF Documents with SAP Intelligent Robotic Process Automation and the Document Translation Service
Machine translation is a commodity service used by many of us every day. Solutions to translate standalone text, websites and even complete Word documents are widely available. However, processing PDF documents challenges most translation services. The Document Translation service powered by SAP Translation Hub currently doesn’t support PDF documents either.
The reason why machine translation for PDF is tricky is the very nature of PDF documents: A nicely crafted PDF usually includes:
- graphics
- structure such as headings, and tables
- formatting such as italics
Simply extracting textual content from PDF and returning nothing but text from translation does not do the trick – you need to transfer elements such as graphics and structural information to the translation result.
For an SAP internal competition – the SAP Enterprise Adoption Low-Code/No-Code Challenge – the SAP Translation Hub team approached me with the question if SAP Intelligent RPA can help here. Sure, it can!
Since Intelligent RPA can automate e.g., Microsoft Office applications, we can use Microsoft Word to open PDF documents, and save them to .docx – or the other way around – in a matter of seconds.
So, together with many other SAP teams we competed for the fame of being the best “No-Coders”.
In our easy to use, low code/no code development environment, I created a small demo bot which does the following:
- Allow a user to select a PDF file
- Open the PDF file in Word and save it as .docx
- Call the Document Translation service to translate the .docx file
- Open the translated .docx file in Word and save it as PDF
Below you can see an example document in Japanese and the translation to English: structural elements like headings, graphics, and formatting are preserved.
Japanese File |
![]() English Translation |
You can see a full demo video of the bot here:
This example shows very nicely how you can use SAP Intelligent RPA to combine desktop applications and cloud services to build an end-to-end solution.
If you are interested in building this bot, please check out the document translation example in the SAP Intelligent RPA store:
https://store.irpa.cfapps.eu10.hana.ondemand.com/#/package/66e5b08a-ec49-4ef0-800f-2542f68af048
(Please note that this bot only works for .docx files – the conversion to PDF is not included).
And if you wonder how our team did in the competition:
We won the first price in the category Internal Process Optimisation 😃.
Thanks a lot to the SAP Translation Hub team for the idea and their help.
Special thanks to: Beatrice Pasch, Christian Lieske and Michael Koehler
For further information about the Document Translation service check out this site:
https://help.sap.com/viewer/9f73362817cd48339dd8a6acba160f7f/LATEST/en-US/a2dedd7861624a1a82d7ec7ea431a8e4.html
SAP Intelligent RPA and related links:
- SAP Community Page | Q&A
- Explore Successful Use Cases | Best Practices
- Try SAP Intelligent RPA for Free Trial Version | Pre-built Bots
- Hyperautomate processes in Finance | Procurement | Human Resources
- Follow us on LinkedIn | Twitter | YouTube
Learn more Help Portal | openSAP | Product Documentation
Very well done, kudos Dietmar Steinbichler, Beatrice Pasch, Michael Koehler and Christian Lieske!