Toolbox – by Shipra Mitra, Principal, Intelligent Automation Consulting, EPAM
With the evolution of intelligent automation, companies still struggle with adoption, starting with getting buy-in from leadership, knowing where to start, setting realistic goals, evaluating RPA vendors and maintaining compliance. EPAM’s Shipra Mitra shares how to address these challenges.
A few years ago, when intelligent automation (IA) and robotic process automation (RPA) began generating buzz, much of the hype focused on bots taking over human jobs. Product companies started positioning the technology as “macro on steroids,” where every process and task could be completed perfectly by a bot. Soon, the industry realized that maintaining these bots was a challenge, and product companies focused on industry-grade RPA tools and platforms. Automating complete workflows became essential. Intelligent document processing (IDP) enabled unstructured and structured data to be classified and extracted, but humans were still required for data evaluation. From there, supervised learning became part of the RPA/IA toolbox. With microservices, API integration, machine learning and IDP, large parts of business processes have the potential for automation in a different scenario than the “macro on steroids” expectation.
With the evolution of IA technology, companies still struggle with adoption, starting with getting buy-in from leadership, knowing where to start, setting realistic goals, evaluating vendors and maintaining compliance.
The first step in adoption is making the case for investment. IA can be transformative, but only if it is adopted by the organization. First, focus on the influencer and the adoption strategy. Early on, find a champion who will drive the initiative. Training, messaging and communication is key for adoption. Positioning IA as a new way of working or increasing efficiency and auditability will gain more buy-in than if it is positioned as bots completing mundane work.
Often, IT departments can clearly see the benefits of automation to assist process owners. When the time comes to implement the bot and move it to operations, business users become apprehensive about the changes that impact the way they work. Business process owners must buy into this new way of working, as they are ultimately responsible for the bot and the virtual task force will “report” to them.
Typically, the security, risk and compliance team is the first to question the disruption bots can create if they are not used judiciously and within standards. The support of the security, risk and compliance team, IT support team as well as the infrastructure team is required to deploy the virtual taskforce, and for success of the RPA initiative overall. We’ll explore more on this later.
Knowing When & Where to Start
Often, an in-depth analysis of technology requirements is not necessary for adoption, as IA operates on current state. There are minor changes regarding how the process will be executed when onboarding bots. In fact, if there are anticipated changes and enhancements, it’s best for implementation to begin before the changes so current pain points are alleviated. Since RPA tools are flexible and changes are configurable, showing an example to end users generates better requirements. For a simple process, a working session with the end user and developer can determine the requirements, design and development with a working bot as the output.
As many IA solutions rely on applications, systems and websites managed by third parties, it’s critical to understand when changes are underway for these sources. A major change in the user interface will impact bot operations, and sometimes even minor changes can impact them. Keeping third parties involved helps maintain the bot workforce, as upcoming changes can be anticipated. Remember that some services dissuade bots on their websites, so permission may be necessary.
Deciding where and how to start with adoption should be evaluated based on the hours it could save or the risks that could be avoided. If the effort required for automation is greater than the value realized annually, it may be better to explore other options.
Typically, when businesses look at IA, the initial objective is FTE reduction. However, with maturity and deployment, companies realize that reducing humans involves much more than just deploying bots.
First, 100% process automation is never achieved. There are exceptions to the process, and bots aren’t great with exceptions. Second, applications and systems change, and bots are not adaptive. Third, bots only do part of the work – automating the process from end-to-end requires capabilities that aren’t covered by bots alone. When machine learning is used, decision-making becomes probabilistic, where the top decisions can be right, and humans need to make the final decision. Therefore, a virtual workforce that is not adaptive, innovative or responsible cannot be compared to humans. A virtual workforce can assist humans by making their jobs easier and allowing humans to focus on true value creation with work that is interesting.
Every year, RPA vendors add new features to their product portfolios, such as vision recognition, business process management or bot libraries. The robotic process automation tools and vendors available are incredibly diverse – ranging from free solutions to annual subscriptions starting at $100,000.
With so many options on the market, it’s difficult for companies to choose the right solution for their business. The following questions can help you find an IA provider:
What is your budget? Less costly solutions may be riskier and less effective based on your needs.
Are you choosing a tool for a specific need or broader capability? Certain features are more important than others depending on your need for automation.
What is your risk tolerance? An industry-grade solution is typically less risky than freeware.
How much time do you have to assess vendors? If you have over a month, then evaluate all your options. If not, choose a top provider with out-of-the-Managing Expectations Managing Expectations box solutions.
Do you need an on-premise, cloud-based or managed services solution? If you want to invest in the product license and infrastructure and are risk averse, go with the former. If you are comfortable with the as-a-service cost, choose the latter.
Is a .NET-based solution preferred? Some products are .NET based. If you are a Microsoft shop, you may want to consider a similar environment.
Starting with this foundation will allow you to narrow down your list of options.
Risk & Compliance
As previously mentioned, bots aren’t responsible for their work and cannot be held liable. However, in HR systems, bots need to be tagged as members to access systems and resources, just like any associate. Some companies create a virtual worker profile to address this challenge, but there are a few differences that need to be incorporated. For example, the password expiration policy will create disruption for bots. Additionally, a bot farm can execute the same activities, which means the same user ID and password may be needed.
The benefit to a virtual workforce is all activities are logged, which isn’t possible with a human workforce. This added benefit also reduces company risk, as you can ensure all processes follow protocol and are standardized and auditable. A virtual workforce not only enhances productivity, but it also reduces risk for a company.
We’ve seen quite an evolution since RPA and IA first came on the market, and no one knows for sure where this next-gen technology will take us. What today’s businesses need to keep in mind is that the challenges of adoption don’t compare to the opportunity that RPA and IA bring to a company’s bottom line.