Robotic Process Automation – The Tip of the Iceberg


In a world of Alexa, iRobot and self-driving vehicles, we should be surprised if robots did not make their way into the contact center.  With about 70% of contact center expense being human labor, that huge cost is a natural target for most enterprises.  If you could provide the same level of service to your customers and cut costs, it’s a proverbial “no brainer”.  There’s a groundswell of attention to the potential for Artificial Intelligence (AI) to replace workers in many industries, but Robotic Process Automation (RPA) is starting to do that today in contact centers across the globe.


According to Gartner, RPA is “is designed to mimic the same ‘manual’ path taken through applications by a human using a combination of user interface (UI) interactions.”  In its simplest form, RPA is a mini-software robot that resides on each agent workstation passing data back and forth among application programs in exactly the same way that an agent would do it, relieving the agent of repetitive, time consuming, manual tasks.  A rudimentary example would be eliminating the cut and paste necessary when transferring an insurance applicant’s personal information stored in the Salesforce CRM to an insurance quoting application.  That’s a mind-numbing task for agents, and for an insurance carrier processing thousands of quotes every day, it’s a significant unnecessary labor cost.  RPA offers a win-win for agents and enterprise.  With the single click of a new RPA desktop button, that multi-step process is reduced from 30 seconds to 3 seconds.

RPA solutions, such as Aspect’s Activity Automation, typically use the field-oriented nature of an agent’s screen to capture and write data across applications.  When combined with a graphical drag and drop representation of data elements and desired data flow that incorporates conditional branching, you have a really powerful way to automate the agent desktop and simulate human agent interactions with applications.  Let’s face it, the agent desktops in most organizations are a hodge-podge of applications from different vendors, and the agent provides the “integration”.

Most of the hype about RPA relates to the futuristic AI world of autonomous software robots that completely supplant the human.  However, Gartner makes the distinction between “attended” RPA, where these mini software robots operate at the behest of agents, and “unattended” RPA, where these software robots are sentient enough to make important decisions that would typically require the judgment of a human being.  Unattended RPA can be implemented on a virtual desktop, or directly through APIs to applications running on a server.  For example, an RPA bot could ingest all the application information for an insurance policy from desktop apps, use pre-programmed rules to make the decision to issue the policy, and initiate the communication with the applicant that the policy has been approved.  Some vendors tout RPA as the agent’s savior, but it will certainly reduce the number of agents required in many back office and front office tasks.

Overlay the rich data access available from desktop apps using RPA with Natural Language Processing and with true AI, and you are ready to turn the lights out.  You have all the ingredients of a human agent decomposed into technology components.  This new “digital employee” can fully emulate the agent’s ability to understand the semantics of the customer request (NLP), has access to all the relevant information and applications (RPA), has the intelligence to gather more information from the customer and make decisions (AI) as well as take action (RPA).  All this technology exists right now except for AI advanced enough to make complex human decisions.

But wait there’s more.  With decreased numbers of agents, there is decreased need for supervisors, quality mentors, workforce planners and administrators.  Each of these managerial positions also relies heavily on WFO and other application programs, and much of the responsibilities of these positions can also be automated with AI trained in each respective area of expertise.  For example, a workforce planner that each day compares actual current call volumes with actual numbers of staff and then issues appropriate requests for voluntary overtime or time off, could be replaced by digital employee technology trained on the heuristics that the workforce planner is currently using to make these decisions.  Aspect Workforce Management already issues these requests using rules-based decisioning, but you need AI to match the sophisticated decision-making of the experienced workforce planner.

Robotic Process Automation of repetitive agent tasks is just the tip of the iceberg that’s floating dangerously close to the traditional contact center.  All the technologies, other than true AI, are already in place to dramatically change what we have known historically as a “contact center”.  Get ready for a wild ride over the next few years.

Latest posts by Mike Bourke, SVP & GM Workforce Optimization (see all)