How Robotic Process Automation and Artificial intelligence can merge for another level of automation ?

[former_member659240]

This is a submission for the SAP Intelligent RPA Tutorial Challenge

This blog post is the result of my own analysis and experience. Do not hesitate to add or correct details by comments.

Please enjoy and feel free to like and share

 

 

In this blog post, I’m talking about how we can implement a logic that brings together artificial intelligence, cognitive automation and robot process automation to create a smart e2e business process then we can reach a cost saving and customer satisfaction.

Before starting, I would like to introduce what we consider Intelligent Robotics and its impact on software development, namely the conjunction between AI and RPA.

 



On this distribution, we can consider that the automation is done on several levels starting from the middle layers platforms to the high level automation with the Box RPA, IA.

The process is unconditionally linked between the different levels for this you must understand the different systems so as not to work twice in the same area and create an intelligent innovation strategy.

Consequently, the challenge in this blog is to mount the different studies to combine RPA: SAP Intelligent RPA or UIPATH (to know that this solution also largely supports SAP transactions)  with Machine learning  to integrate, automate and run an independent robot.

 

 

Breaking it down

 

Robotic Process Automation is the game-changer that can run on central server without using any technical environment (no human hand).

RPA Bots can be triggered in attended mode, intended mode and via an api call helping us to automate manual tasks on a large scale in a very secure environment.

Before putting the link with artificial intelligence,I would like to emphasize the fact that AI has made enormous progress. My first interaction with this field was summarized on image and sound processing on Matlab, then I started on a different python programming environment concentrated in the Anaconda distribution: Prompt anaconda, Spyder and the one I prefer the most Jupyter notebook.

Recently I discovered a powerful service of SAP Leonardo machine learning which is called SAP Data intelligence and which supports a flexible development IDE with possibility of installing all the libraries necessary for data processing.

Here we see the next step is coming. It’s mostly cloud-based solution and the range tasks that this service can handle is much bigger than what can expect today.

RPA today can replace the hands perfectly when Machine learning in the future will be able to replace cognitive abilities using Natural Language Processing, computer vision and cognitive computing.

 



 

 

Proof of concept

 

Use case 1:

 



 

On this use case, we can see that the model runs an automation flow from one system to another, the different systems, websites for example have security platforms (e.g: Captcha Iframes).
To solve this problem, we develop a model that calissifies and recognizes captcha (Images, digits and so on ).

After creation and training of our ML model on Jupyter notebook, we can deploy it in a precise environment using the service modeler : the orchestrator which is based on pipeline engine that contains several workflows.

Once the model is deployed a REST API service is generated, with this option we can call our model in the Robot workflow using the API service provided by the RPA solution.

**************************************************************************************************************

 

Use case 2:

 



 

In this use case, we have three different systems with a constant data flow.
the first entry of instructed data goes through the machine learning service in the cloud.

After processing the data is more structured and ready to be scrapped by the robot which will then do activities on another system.

 

Conclusion

we consider that a perfect understanding of the different capabilities of RPA and machine learning leads to innovation with intelligent and fully embedded systems (100% automatable).

 

Let me know your feedback in comments.

 

EL MARHNAOUI SALAH

 

(All images are created by the author himself)