What is the difference between cognitive RPA and RPA?

I would consider the addition of AI/ML to your RPA to be the most important Cognitive RPA feature. 

E.g., UiPath AI Fabric allows you to consume information from AI/ML and use the result in logical decisions and to inform human teammates. You can set up a feedback loop to continue training your model to improve efficiency and confidence.

Let's say you're willing to automate a 95% confident decision without human intervention. Anything under that confidence you'd route to a human. You can use the results of the human decision to teach the model and make the AI/ML even more effective. This allows your AI/ML to get smarter and ever more helpful (cognitive) as time goes on. 

Cognitive RPA enables you to design more complex and less rule-based processes using AI-powered bots integrated with third-party cognitive services, mainly from Google and Microsoft.

The RPA and AI technologies collaboration provide automation capabilities that level up business values. These unique capabilities include optical character recognition (OCR), text analytics, document processing, content recognition, machine learning, etc.

Using the cognitive functions of RPA extends the opportunities of process automation, enabling intelligent bots to make decisions while learning the role as they become valuable sources in your digital workforce.

-Reducing operational costs and barriers: RPA can automate your legacy systems and connect your operations with AI technologies, unblocking all restrictions.

-Arrange and process unstructured data: RPA bots with cognitive automation capabilities capture and manipulate data from unstructured sources, making the data structured in a way you need. 

-Automation of business processes: With RPA Cognitive Automation, you can save time and money by automating any business process, reducing human interaction, and increasing productivity.

-End-to-end customer experience: Increase customer happiness by providing quicker response times, higher accuracy, and consistent results. 

Hi Evgeny,

Cognitive AI we develop includes the skills to percept and to control anything on a screen in a similar way, a human employee does. A Cognitive AI is independent of the position of objects and can even act flexibly on the appearance of such objects. 

Benefits: 

1.) it is very easy to tell it, what to do and how to do it since I can "tell" her, how I am operating the processes and she will do it in the same way. The consequence is that also employees without an IT background can automate processes. Cognitive AI is a no-code way to automate. Low-code solutions by definition still require coding skills and only speed up a developer.

2.) It is very fast to tell it how to operate a process because I don't have to deal with complex technical selectors and activities.

3.) For Cognitive AI it does not make a difference, if it looks through cameras in order to operate physical devices such as tablets, smartphones or machines or if it just "looks" at the perfectly rendered screen of your desktop in order to automate PC applications. 

4.) Cognitive AI does not require the object IDs ordinary RPA solutions use, so the processes that have been automated do not stop when the object IDs change - e.g. very common after system updates. The consequence in "old school" RPA projects is, that a full DevOps team of developers is necessary to keep already automated processes running... A very expensive consequence of automating the process of an employee with a moderate salary in order to transfer the know-how into a bot that needs to be maintained by software developers.

And since I am a software developer myself, I tell you something: we don't like this kind of work. We love to develop and evolve new and powerful systems rather than maintain IDs in broken processes...

And there is still enough stuff to automate - the industries are still at the beginning of the digital transformation.

Please feel free to DM me for more details about Cognitive AI. 

Kind regards,

Michael

Marketing...! 

This is not to say that there isn't value in combining machine learning and RPA to enhance the values of the solutions you build - there is value there, and we're big proponents of building AI and ML into the solutions we deploy.  

But let's not confuse machine learning with cognition, and I think we're all better served just focusing on the specific problems we're solving and how rather than how the sausage is made in terms of the tools we use.