Insights & Use Cases

What We Offer

Strategy

We help you develop a Strategy and a Roadmap, including Tool Selection, Productivity metrics and ROI Models.

Implementation

We deploy your RPA solution, establish governance and address IT and security concerns. We also train your staff to maintain the solution.

Managed Services

We manage, monitor, tune and continually optimize your robotic process execution. We also implement enhancements and manage your RPA infrastructure.

TALK TO US

Artificial Intelligence and Process Automation

From the beginning days of IT, one of the key goals has always been to Automate Manual Business Processes. From Mainframes to Client Server Application Development, to Internet, Business Process Management(BPM) based Workflows and Rules Processing Technologies, we have come a long way. As we developed Applications to automate business processes, we realized that most of the Applications end up being siloed solutions for specific business needs. We used Enterprise Integrations to tie things up and started using BPM tools to re-engineer the entire spectrum of Business Processes using Workflow tools and Rules Engines.

With each advancement though, the need to address tactical Automation challenges continued to evolve as well. Macros, Scripting, Screen/Web Scraping type of Technologies have been in use for a long time to address such challenges. Go back a few years when the RPA evolution, all the tactical Automation technologies were organized under an Enterprise RPA Tool umbrella; whereby you could combine the power of Surface/UI level Automation using Screen Scraping, Macros, Scripts, APIs etc. These Automation “Robots” are now capable of manipulating pretty much any type of Application that exist out there including Mainframe, Web, and legacy Client/Server technologies. The Robots can be deployed and tasked via a Control Tower or Orchestration Console and monitored by the Digital Workforce.

The traditional RPA Robots by and large are “dumb” technologies because they are unable to make subjective decisions. In addition, most of the RPA tools rely on Structured data inputs and can’t process Unstructured Data such as OCR images, unstructured text (emails etc) etc. At a high level, a Robot starts to become “Smart” if it can.

  • Process Unstructured data by extracting relevant Data elements from it. (For e.g. an unstructured email paragraph which contains some pieces of Data but is not in a pre-defined format)
  • Make some subjective decisions based on specific set of data. For e.g. determine if a given address is that of a Company or an Individual (“John Doe” vs. “John Doe Inc” or “John Doe LLC”)

Case Study

Let’s take an example of a simple Business Process where someone in Accounting Department must manage the Aging Accounts Receivable problem. The User pulls up AR Aging report, pulls up the relevant Invoice information and sends the Invoice and an Overdue notice to the Customer. So far this can be accomplished by “dumb” Automation. However, if the Business Process requires that the Customers who have a pattern of paying late, need to be handled differently, then the Robot needs to be able to identify that pattern just like the Human would. It’s quite possible that pattern could be identified by a set of rules however in some cases there is subjectivity. For example, that Customer may be going through financial troubles and there is some outside news or information which could identify such pattern. This is where a smart Robot would not only look at the AR information but also look at payment patterns as well as analyze the public news or “sentiment” on the Customer and flag it as a case which needs to be handled differently.

These Smart Robots are driven by Machine Learning Technology, so it is important to do a deep dive during your RPA vendor comparison to understand how you want to enable it.  Some RPA Vendors have these smart functionalities built in whereas others open connections into their platform so these capabilities can be plugged in as the Process Automation is developed. In the example above, the Robot would not only look at the Customer payment history but would also use public information about the Customer using “Sentiment Analysis” AI Models to flag if a Customer has Financial issues.

Conclusion

There are endless use cases, another example would be a Desktop Robot assisting a Customer Service Rep.  By analyzing Customer Complaint emails for negative or positive Sentiment as well as extracting Customer Account Number, the robot is then able to pull up the holistic view of the Customer in front of the CSR without the CSR going through many different Screens and Applications. The end result is a better informed CSR and a better Average Handle Time for a Customer Complaint.

The Smart Robot revolution is upon us and its happening much faster than we think. The good news is that AI Technology is fast approaching a point where a lot of us can consume these Machine Learning Models without concerning ourselves too much about the inner workings of how the AI Model is developed. The AI Models do have to be trained using Training Data and validated using the Test Data.

RP eBook Post