What is Cognitive Process Automation?

February 26, 2021
Conceptual image of coding inside of a dark box.

Remember when everyone started slapping the label “smart” on everything? From phones to water and the wide range in between, it became a way to make a regular product sound fancy. And let’s face it, it worked. Somehow when something is labeled well, it seems more impressive. More exciting. More …smart.


The same is true with Robotic Process Automation (also referred to as RPA). The phrase conjures up images of shiny metal robots carrying out complex tasks. Especially if you’re not intimately familiar with the tech industry and its automated contributors, Robotic Process Automation probably sounds impressive. Exciting. Smart.

This was once true, absolutely. What we know today as Robotic Process Automation was once the raw, bleeding edge of technology. Compared to computers that could do, well, nothing on their own, tech that could operate on its own, firing off processes and organizing of its own accord, was the height of sophistication. However, that this was only the start in an ever-changing evolution of business process automation.

Types of Business Process Automation


Although there’s an alphabet soup worth of acronyms to look at in the automation space, they all come down to the same thing: Business Process Automation (BPA). To put it bluntly, this is the concept of automating the mundane, tedious, and repetitive work.

Although the evolution of BPA has changed, the foundation has always been the same: get your team focused on strategy, growth, and customer satisfaction. It would be easy to assume that large companies with enormous budgets would have automation nailed down; however, the reality is these organizations are some of the least streamlined organizations. The more processes, systems, and headcount in a company, the harder it is to transform it into a lean and mean machine.

As business leaders around the globe have recognized the need for dramatic transformation, they are not looking for dramatic company disruption. Innovation has helped ease the pain of implementing automation and getting the workforce back to the root of what they’re trying to accomplish.

Generation 1: Robotic Process Automation


The simplest form of BPA to describe, although not the easiest to implement, is Robotic Process Automation (RPA). This first generation of automation, when emerging, was the pinnacle of sophistication and automation. It created the foundation for the future evolution of streamlining organizations.

The Benefits to Robotic Process Automation


RPA is a phenomenal method for automating structure, low-complexity, high-volume tasks. It can take the burden of simple data entry off your team, leading to improved employee satisfaction and engagement.


The Downfall of Robotic Process Automation


Like any first-generation technology, RPA alone has significant limitations. It is built based on “if this, then that” statements. The business logic required to create a decision tree is complex, technical, and time-consuming. In addition, if data is incorrect, unstructured, or blank, RPA breaks. Your team has to correct the system, finish the process themselves, and wait for the next breakage.

Another key issue with traditional RPA is the cost. Organizations with millions in their innovation budget can build or outsource the technical expertise required to automate each individual process in an organization. It can take anywhere from 9-12 months to automate one process and only works if the process and business logic stays the exact same. Even a minor change will require massive development and testing costs.

Practical Application of Robotic Process Automation


Think about the incredible amount of data flow running through a financial services company for a moment. As companies are becoming more digital daily, we will use the example of a structured, accurate, online form. The assumption is all the data is valid and correctly filled out.

As an example, you have an insurance policyholder that wants to file a claim online. The structured data in that form can be send to a Claims Adjuster, filed into the claims system, and fill out any digital documentation required. This eliminates much of the manual work required by a Claims Assistant.

Generation 2: Intelligent Process Automation


The next breed of Business Process Automation is Intelligent Process Automation (IPA). Exactly as it sounds, it is the concept of injecting intelligent, machine learning capabilities into Robotic Process Automation. This amplifies the capabilities of automation from simply “if this, then that” into more complex applications.

The Benefits of Intelligent Process Automation


Although Artificial Intelligence (AI) is often used as a buzzword in the technology space, it’s commonly misused. Rather, it is Machine Learning (ML) that is used in Intelligent Process Automation that makes it, well… intelligent. Machine Learning helps Robotic Process Automation recognize patterns and improve through experience. The more data you provide the technology, the more it learns. This enables Intelligent Process Automation to take on more complex and advanced processes than Robotic Process Automation alone.

Inherent Challenges of Intelligent Process Automation


Although Intelligent Process Automation leverages Machine Learning to avoid mistakes and breaks in the system, it has some of the same issues as traditional Robotic Process Automation. First, it is expensive and out of reach for most mid-market and even many enterprise organizations. The setup of an IPA algorithm and technology requires several million dollars and well over a year of development time in most cases.

The expertise required is large, and although you can outsource it, the algorithms require vast amounts of maintenance and change management. Any system, process, or technology changes requires a great deal of development.

Practical Application of Intelligent Process Automation


For those that can reach the cost and timelines required of Intelligent Process Automation, there are a great deal of applications within reach that exceed the capabilities of “if this, then that” statements alone. Let us revisit an insurance application again. While Robotic Process Automation is not able to read documents, Intelligent Process Automation gets us started down this path.
Take a standard CMS 1500 Claim form as an example. The way Machine Learning works is you create a “mask” over the document that tells the algorithm where to read specific pieces of information. This information can then be picked up by the Machine Learning and continue down the path of entering the data into systems, alerting a Claims Adjuster, etc.

Generation 3: Cognitive Process Automation (Digital Coworkers)


The newest, emerging field of Business Process Automation lies within Cognitive Process Automation (CPA). While Machine Learning can improve algorithms, true Artificial Intelligence can make inferences, assumptions, and teach itself from abstract data. This is the concept that Cognitive Process Automation is built on. It solves the issue of requiring extremely large data sets, budgets, maintenance, and timelines that only innovative, enterprise organizations can afford.

Human Intelligence, Automated Speed


Artificial Intelligence replicates the intelligence of a human. Traditional automation requires clear business rules, processes, and structure; however, traditional manpower requires none of these. Humans can make inferences, understand abstract data, and make decisions. If you change variables on a human’s workflow, the individual will adapt and accommodate with little to not training. Cognitive Process Automation brings this level of intelligence to the table while keeping the speed of computing power.

Practical Application of Cognitive Process Automation


Where do you tend to use humans instead of machines? It’s typically where documentation, decision-making, and processes aren’t clearly defined. Going back to the insurance application one last time, think of the claims process. Would you ever let a bot lacking intelligence determine whether a claim is approved? Likely, no.

Cognitive Process Automation learns from observing Claims Adjusters and creates its own algorithms for approving or denying claims. If it isn’t sure what to do, it will ask your team for help, learn why, and then continue with the process as seamlessly as a human. This level of technology can even help Underwriting teams determine straightforward policy administration, Finance manage Accounts Payable, and Human Resources put onboarding and offboarding on autopilot.

The Cost of a Digital Workforce


You might’ve heard of a Digital Workforce before, but it tends to be an abstract, scary idea. Reality is far from it. A Digital Workforce is the concept of self-learning, human-like bots with names and personalities that can be deployed and onboarded like people across an organization with little to no disruption.


A digital workforce, like a human workforce, is pre-trained and ready to work for you. These bots specialize in their field just as an Underwriter, Loan Officer, or Accounts Payable Specialist does. With 80% of their needed knowledge already pre-developed, they can plug-and-play in just a few weeks, teaching itself what it doesn’t know. Since the technology can adjust itself, maintenance is near non-existent. This significantly reduces the costs across every stage of the technology life cycle. Compared to the millions required in RPA and IPA, Cognitive Process Automation can often be implemented for as little as the cost of adding one person to your workforce, but with the output of four to eight headcount.

Meet Your New Digital Coworker


Roots Automation was founded specifically to bring Digital Coworkers to the market at scale and reduce the barrier to entry to insurance, banking, and healthcare organizations around the globe. Created by leaders at AIG and Mars with decades of automation experience under their belt, the Co-Founders of Roots Automation have pre-built, plug-and-play Digital Coworker bots across Finance, Operations, Underwriting, Claims, Human Resources, etc.

You now can streamline and automate your business more efficiently and cost-effectively in a time where every company is striving to get lean and mean. With so many unknowns in the market, profitability and client retention are the goals of nearly every business leader right now. Employ your first Digital Coworker in as little as three weeks and see your break-even point in as little as four months.


Skeptical of Digital Coworkers in your business? Read “The Nail in the ‘I Can’t do Automation’ Coffin”
Want to learn more about Digital Coworkers in your business? Schedule a demo.
Unsure of where to implement a Digital Coworker? Check out our case studies.

Originally published on November 23rd, 2020 and updated on February 26th, 2021.

Demo a Digital Coworker On-Demand

Artificial Intelligence and automation can be difficult to conceptualize.

Access our on-demand demo and watch as a Digital Coworker learns to underwrite an insurance policy, eventually reaching a 95% straight-through processing rate.
Take me to the on-demand demo
Balancing Employee Engagement and Digital Disruption with RPA
Read More >
Cognitive Solutions in Business Efficiency and Performance
Read More >
The importance of standardization in the age of RPA
Read More >

Subscribe to our blog

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.