Both cognitive agents like Amelia and robotic process automations (RPAs) deliver value to enterprises through the automation of transactional tasks. However, only cognitive agents are able to automate tasks related to decision making while delivering an intelligent user experience.
RPAs execute tasks, cognitive agents automate intelligence
As AI grows in prominence, so have several related terms that fall under the greater AI umbrella, e.g. automation, autonomics, machine learning (ML), etc. While these technologies may share features in common, there are also many important differences.
For example, both robotic process automations (RPAs) and cognitive agents like Amelia are able to automate tasks, however only Amelia is able to integrate complex human interactions and human-like analysis. As an article from CIO noted, while RPAs are able to automate rules-based business processes, many observers see them “as a stopgap en route to intelligent automation (IA) via machine learning (ML) and artificial intelligence (AI) tools, which can be trained to make judgments about future outputs.”
RPAs are useful technologies, but advanced cognitive agents are far more powerful. Here are four important ways that an advanced digital colleague like Amelia can deliver more business value than a low-level RPA.
Cognitive agents speak fluent Human
In a recent post, we detailed the differences between cognitive agents like Amelia and mere chatbots — including Amelia’s ability to handle a wide spectrum of human inputs. This is an important differentiation as Amelia can handle queries and commands for which she hasn’t been specifically programmed, whereas a chatbot follows static decision trees and can only successfully navigate restrained interactions.
While both RPAs and cognitive agents automate tasks, RPAs can only accept specific pre-programmed user commands (if they accept any at all), which limits their usefulness.
This versatility is even more important when it comes to differentiating cognitive agents from low-level RPAs. While both RPAs and cognitive agents automate tasks, RPAs can only accept specific pre-programmed user commands (if they accept any at all), which limits their usefulness.
For example, both Amelia and RPAs can reset a user password, but only Amelia can be prompted via a versatile human interface, so a user can simply say or write, “Amelia, will you reset my system password?” More importantly, Amelia is able to accommodate a wide spectrum of human conversational inputs including very casual ones like, “Hey, I forgot my password, can you take care of that for me?” The ability to talk to humans on their terms is important because it lowers the bar of access to all users regardless of technical prowess, thus expanding their utility.
Cognitive agents can make decisions
The same functionality that allows cognitive agents to process a wide spectrum of human inputs also allows them to process data, make predictions, and render decisions. This means that Amelia can independently change workflows in response to relevant data, whereas low-level RPAs follow preset rules and are far less versatile. For example, Amelia might upsell products to a customer in a follow-up email after acquiring specific demographic information through user interaction – a complex workflow that low-level RPAs were not designed to handle.
Cognitive agents can self-learn
RPAs follow pre-programmed rules and never deviate unless a human proactively changes the rules. Cognitive agents, on the other hand, tap into ML functionality to independently improve at tasks over time. For example, whenever Amelia is not able to assist a user, she can elevate the session to a human colleague and observe the interaction in the background, which allows her to expand her knowledge base for future user interactions. To acquire new functionality, an RPA requires human intermediation.
Cognitive agents can bridge enterprise systems
A report from HfS Research recognized Amelia as the “digital glue between front and back offices,” which means she can connect users on the front end to back-end systems through the aforementioned versatile natural language interface. While a low-level RPA can be designed to bridge specific systems for a specific task (e.g. if a retail store’s inventory on a particular product falls below a predetermined threshold, the RPA can automatically order a new shipment), a cognitive agent can connect resources throughout an enterprise to accomplish goals, particularly when part of an autonomic framework such as 1Desk. This allows a user to execute complex actions across a system. Amelia, for example, allows an executive to perform a complex IT task using natural language such as, “Please forward all of my incoming mobile phone calls to my assistant’s desk number while I’m on vacation for the next week, unless it’s a call from the CEO’s office.”
While RPAs can automate rules-based processes featuring transactional tasks, cognitive automation allows machines to take on advanced functionality, which makes operations more efficient and adds business value. Cognitive agents have a greater degree of intelligence compared to RPAs – simply because they were built that way.