Cognitive Automation 101

Share this post:

What is Cognitive Automation

Automation is a fast maturing field even as different organizations are using automation in diverse manner at varied stages of maturity. At the most nascent stage, some enterprises are still using Desktop scripts or Robotic Process Automation (RPA) and deriving value from those while the more mature organizations have moved beyond RPA into what IBM calls Autonomic Process Automation and finally Cognitive Automation. As the maturity of the landscape increases, the applicability widens with significantly greater number of use cases but alongside that, complexity increases too.

At IBM, we believe that true value of Automation can be leveraged when Cognitive meets RPA and enable autonomous decision making, understanding natural language, self learning and ability to handle scenarios that entail unstructured data and complex decision making. For example, Automating a process to create a support ticket when a database size runs over is easy and all it needs is a simple script that can check the DB frequently and when needed, log in to the ticketing tool to generate a ticket that a human can act on. However, if the same process needs to be taken to logical conclusion (i.e. restoring the DB and ensuring continued business operations) and the workflow is not necessarily straight-forward, the automation tool-set needs to be expanded heavily. In most scenarios, organizations can only generate meaningful savings if the last mile of such processes can be handled .

IBM’s Cognitive Automation is a powerful augmentation of standard RPA capabilities that allows for an end to end execution of automation requirements even in following situations –

1. When data is unstructured – such as making sense of email text
2. When non-standard documents are to be handled
3. When decision rules are complex, and
4. When interactions are required with multiple disparate systems

IBM’s cognitive Automation Platform is a Cloud based PaaS solution that enables Cognitive conversation with application users or automated alerts to understand a problem and get it resolved. It is made up of two distinct Automation areas; Cognitive Automation and Dynamic Automation. These are integrated by the IBM Integration Layer (Golden Bridge) which acts as the ‘glue’ between the two.

Think of it as a combination of Brain , Nervous System and Hands & Legs. Cognitive component provides the “Thinking” layer that provides the understanding of the context and ability to decipher the right action . The nervous system is the orchestration layer that carries the instructions from the Brain to the Hands and Legs. Those hands and legs provide the end point execution of tasks as directed by the brain. Consider a scenario where email based instructions require half a dozen tasks to be performed depending on nature of the instruction. And those emails are written in free form by different end users. Cognitive component can make sense of those emails and provides the instructions to the integration layer , that is Golden Bridge , which in turn would invoke the right robot to perform the actual action. This architecture not only allows for end to end automation but also provides framework for rapid scaling and effective utilization of costly robots [because the robots are invoked only when they are required].

Thus, Cognitive Automation can not only deliver significantly higher efficiency by automating processes end to end but also expand the horizon of automation by enabling many more use-cases that are not feasible with standard automation capability.