As part of Datalabs offerings, were evaluating multiple platforms for Robotic Process Automation (RPA) and we explored few custom development options as well and there were a lot of interesting learnings.
As part of custom development, we did an RPA platform to optimize an online stock trading algorithm (https://www.youtube.com/watch?v=5Xk4AWKPYNY&t=38s) and maximize its performance. This Fintech ML platform yielded excellent results in real-time trading. Besides, we did another custom RPA development (https://www.youtube.com/watch?v=3CQyBZOEmvo&t=107s) where we used an NLP processor to extract instruction from English text to run an automation script to send ‘job offer letters’ to multiple people.
On the platform front, we evaluated multiple options and narrowed down UiPath and we have completed a couple of PoC implementations with real-time use cases. Few take away:
A. Custom development using Python is a good starting point to test the waters and validate your process automation goals. Python-based automation with a cognitive ML backend can solve your process automation challenges. The key requirement is the availability of data sets to coach the platform and well-defined problem statements.
B. When it comes to platforms, we went with UiPath for the following reasons:
1. Simple to use and easy to train resources
2. We always look for community support while choosing frameworks as it would help clients in managing their solutions in long run with better availability of resources. UiPath has a good resource pool with basic certification
3. UiPath fairs better when it comes to OCR and image recognition
4. UiPath is tightly integrated with MS Dot Net and resources with good SQL skills and Dot Net framework understanding fair well with the implementation. We have trained our Microsoft.NET developers with good SQL skills on the UiPath platform and that seems to be an exciting recipe.
