Robotic process automation (RPA) is gaining enterprise acceptance, as the C-suite welcomes the concept of reduced costs and improved productivity through automation. RPA’s foundational technology consists of an army of intelligent software robots that mimic the actions of real people and eliminate manual, repetitive work by humans—cascading positive impact on the bottom line. By leveraging the automation power of a robot, organizations not only make their processes run more efficiently but also enable business groups to monitor processes over time and learn where further improvements can be made.
Deploying RPA can improve enterprise performance quickly, as well as set the stage for increased levels of automation in the future. To ensure both current and future needs are part of the decision, make sure you address key considerations that could create long-term impact.
Begin With the End in Mind
Start by examining the reasons for choosing RPA in the first place. The problems that can be solved by this technology may differ from enterprise to enterprise, but all organizations can benefit by first seeking out hidden process costs within the organization. But hidden costs are, well, hidden, so this is not necessarily a simple task. The trick is to be willing to uncover and embrace broader problems that stem from manual data entry, such as generally inefficient operations, poor customer service, failure to comply with regulations, or the inability to go to market faster than competitors.
- Are your networks of internal and external systems inefficient or too often idle? According to the Cognizant Center for the Future of Work, organizations automate just 25–40% of their workflows. Meanwhile, employees spend 22% of their time on mundane, repetitive tasks.
- Do you lose customers or deals due to slow or repetitive processes? Whether you’re onboarding a new customer with manual processing of data and documents, or providing customer service that involves the collection, integration, and input of information into a business process, performing these tasks manually reduces customer satisfaction and causes customer churn.
- Are you spending increasing amounts on either compliance controls or fines from lack of compliance controls? Deutsche Bank was recently fined a record £163 million for failing to maintain an adequate anti-money laundering (AML) control framework for several years, and it is not an anomaly. Humans make 10 errors for every 100 steps of a manual process, a key reason for enterprises to consider technologies such as RPA.
- Can automation improve the speed and accuracy of your operations, pushing your human staff into other, more strategic roles? Humans perform manual labor work 8 hours a day, need lunch breaks and vacations, and sometimes make mistakes that create other hidden costs such as compliance and customer service issues.
RPA should be implemented before an enterprise realizes it simply can’t keep up with competition that has already streamlined and automated operations. But, what works for a competitor may not be the right path for everyone in the space. Enterprise RPA deployments can and should vary from organization to organization based on goals and objectives.
Are We Really Talking RPA?
It’s important to understand whether the solution under consideration is truly RPA or merely a re-tooling of existing software to capitalize on the RPA boom.
Deployment must allow for software robots to be built in a way that truly mimics what a user does in an application; without this focus on automating routine activities, functionality could be severely limited, restricted to select business use cases, or bolted onto another unnecessary solution that adds cost and time to rollout. To keep data initiatives in check, keep solutions focused on key integrations only, such as enterprise capture, analytics, or process intelligence, that may be critical to the initial RPA deployment itself.
Are the Coding Demands Significant?
While some of today’s RPA options require coding—which will ultimately necessitate a developer skillset and a steeper learning curve—the design of robotic processes, both simple and complex, should be approached in a visual and highly intuitive manner that eases the design of processes and avoids the need for coding, a developer background, and weeks of training. An optimized design studio relies on a model of reusable components, enabling parts of one robotic process to be reused in many other robotic processes. This saves time and makes it much easier to maintain robots when the enterprise really begins to ramp up deployment.
What’s the Spend on Support and Services?
Licensing and training strategies are critical to the initial rollout and success of a project, but it is important to not end up in a trap in which the enterprise ends up paying more for services and support than for the actual software. Consider instead that RPA is designed to leverage professional services and solution integrators to build an RPA foundation, enabling the enterprise’s team to own RPA going forward. Becoming proficient should be a rapid process based on a reasonable amount of professional services rather than a long-term contract, with onsite or online training helping enterprise users design, test, and deploy robots quickly.
What Defines Enterprise-Ready RPA?
Enterprise RPA platforms are designed from the get-go as server-based architectures. Yet many RPA architectures started out as macro recording desktop tools and have since scrambled to accommodate robotic processes that are deployed, managed, and run more effectively from a centralized server model. This may sound obvious, but the reality is most RPA technologies require all of their bots be deployed to a virtual desktop environment in order to run the bot process. In contrast, a true enterprise architecture—with a built-in browser engine on the server and native connectivity to mainframe terminals—does not require connectivity to a virtual desktop for web or mainframe bots. This significantly reduces hardware and software licensing costs as well as IT resources for setting up and maintaining a virtual desktop infrastructure. Additionally, a single robotic process can be run in parallel many times on a server, ultimately delivering superior performance and scalability.
Isn’t Artificial Intelligence (AI) the First Step?
Diving into AI without a plan (and making the discussion all about advanced learning) could result in many months, or even years, of deployment and large software licensing investments that don’t deliver clear ROI. Instead, address obvious process challenges first and include a discussion of where cognitive computing has the potential to solve both current and potential business problems. Where and how does AI fit? It’s a great question; however, CIOs can get early wins with RPA and never touch advanced learning technology. Advanced capture is often all that’s really required for initial data extraction and document classification.
A Shift in Approach for IT and a Bonus for Business
When it comes to software robots, RPA may be the beginning of something big, laying the groundwork for content analytics, natural processing language, and process intelligence.
As an emerging technology, RPA can solve the challenges organizations face with automating processes that involve legacy systems while also dealing with the growing number of external data sources (websites and portals) and desktop applications (e.g., Microsoft Excel and email) integral to today’s enterprise. It’s no simple feat to connect all these applications and data sources.
Traditional business process automation approaches can be expensive and take months or even years to roll out, and no IT department has unlimited time, budget, and resources to address these needs. This is often referred to as the “long tail” of automation, where RPA is focused on the process activities that have historically not been addressed by IT. Asking the right questions will help develop clear expectations for these initiatives—enabling a better understanding of how RPA can position an enterprise for long-term competitive advantages through automation.
For success, solutions must be deployed with both immediate impact and the future in mind. For enterprise innovators, it is critical to understand not only how RPA solutions are defined and distinguished from one another, but also how they can steer your organization on a long-term competitive course featuring continued automation in analytics and learning.
- Not all automation software qualifies as RPA. Don’t buy into re-tooled existing applications that come with unnecessary add-ons.
- One software does not likely do it all. If that is the claim, beware of third-party resources layered in and adding complexity and cost.
- Designing robotic processes should be visual and highly intuitive, regardless of skill level of the designer.
- Validate the software license structure and fees, which should not include large costs to build robots.
- RPA training should not require months of professional services; the concept is designed to be supported with rapid onsite or online training.
- AI is not required as step one. Instead start by addressing obvious process challenges first, based on a discussion of when, how, and if AI should be deployed as well.