Keys to a Successful RPA Implementation in Manufacturing Automation


Many manufacturers have started to explore RPA implementation as a means to automate work processes for various departments. The business benefits from RPA (Robotic Process Automation) come in several forms; cost savings, accuracy, throughput, perfect compliance to regulations, matching audit requirements, and so on. Although the number of companies that have started RPA journeys has exponentially increased, there are only a handful that have been able to fully realize the benefits of RPA through successful implementations.

A successful robotic process implementation is labeled as such when teams have established a repeatable methodology and practice which results in the production/automation of hundreds of tasks or processes. Organizations can deliver a successful RPA implementation process by adopting a roadmap that leads to a sustainable practice. The RPA implementation steps can be broadly divided into 3 phases - incubate, build, and optimize.

Incubation Phase

Most companies start their automation journey with the development of proof of concept automations. This usually starts at a function level by someone who may not be a sponsor or have a mandate at an organization or department level. The scope of the POC is also not well defined and usually entails the automation of some small parts of a larger process that was considered good for automation based on subjective criterion. In most cases, the identified scope is then self-implemented after completing some basic initial trainings on variou RPA technology platforms.

These organizations struggle after the proof of concept, since there is no clear path for scaled adoption defined. They often turn to a search for additional processes and implement those. However, after automating 4-5 processes, they are unable to realize the benefits and the RPA Journey for most ends there.

Although it often results in program failure, this initial phase is very critical and is usually the foundation for setting up a successful adoption of RPA. The difference is that it’s important to not just get any start, but to get the right start. Before implementing RPA, the first question any organization should ask is “What can RPA do for the Business?” as opposed to “What can we do with RPA?”. Answering this question builds the right mindset and is a critical step in the right direction.


Build Phase

Automation, when executed properly, can provide an enterprise with endless benefits. The biggest automation challenge firms face today is identifying a large enough quantity of appropriate workflows fit for automation. After identifying and automating the most obvious automation candidates, the business hits a brick wall and is unable to carry forward the automation momentum.

To solve this problem, enterprises should build an iterative methodology to continuously review processes. These activities should not only provide a list of automation candidates as an output but also a provide a thorough recommendation of overall task reassignment and
prioritization to make operations more efficient.

A structured way to accomplish this is by creating a map of the entire organization to identify the various departments first. Then, in an
iterative way, work with the various department leads to create a map of their department teams along with a list of team members and the tasks they perform. Next, you will be able to evaluate each task to understand if it’s suitable for automation. Each task should then be categorized into different buckets. Some tasks are mechanical (do not require a human decision) and repeatable, and some are Intelligent, where human judgement or a decision is required. There will also be some tasks that have no organizational use or are inefficient. The idea would be to automate the manual tasks, leave the Intelligent tasks for humans (or use AI and other technologies as a next evolution to organizational automation, further described in Optimize phase) and eliminate the unwanted tasks to increase efficiency by rearchitecting the inefficient workflows.

Optimize Phase


During the Build Phase, a mature automation practice should have a process discovery methodology in place to identify processes and categorize them into three buckets - Mechanical, Intelligent, and Wasteful tasks.

Until now, the enterprise automation focus was on tasks that were strictly rules based and leveraged structured data sources when human interpretation was not required. These tasks were easier to automate. For example, it is easy for the system to extract financial data when organized in rows and columns within a table, but extraction of the same information when presented in a multi-page financial statement becomes more complex and requires intelligence for contextual understanding. A use case of this; A home appraisal company is required to evaluate the condition of a house as part of the appraisal process and an appraiser must physically look at pictures of the house to label the property in “Good”, “Fair”, or “Bad” condition. During the build phase we left such tasks for humans to perform, but, in hopes of further optimization, an enterprise should now fit these tasks into automations. Upskilling will be critical to maintaining a skilled workforce that can drive the manufacturer forward.

Automation of Cognitive functionality is the next logical step in building a scaled RPA program as it helps in realizing the true value of RPA. Cognitive functionality automations can also help improve the company’s overall customer experience by providing them with a seamless, fully connected workflow.


Migrating processes to an automated environment can also create additional opportunities, such as generating operational insights. When a human executes a task, there is limited capture of the data trail. An automation environment should have a data collection component that receives and stores the information and KPI data relayed by bots.

For example, an accounts receivable process that sends emails to customers notifying them of overdue payments, may asynchronously log delinquency information as it completes the transaction. Such data trails can then be used for gaining insight and further analyzation.


As you start building new automations the need for infrastructure and platform licenses increases. Although the platform may be shared between departments and managed centrally, if not controlled, it can lead to an overall increase in infrastructure and license cost. Also, this additional need for infrastructure and bot licenses creates a need for an increased maintenance and support capacity.

Inevitably, increased infrastructure capacity cannot be fully avoided, however, it should not be at the same rate at which the number of automations is increasing. Enterprises should ensure that the bots running automations are efficiently configured and executed. One way to do this is by building an intelligent scheduler to increase bot utilization by applying defragmentation techniques to fill bot idle time by relocating the task execution times of other bots.


It’s critical to have a real time view of the impacts caused by changes in applications or other dependencies. A mature automation practice should plan for this capability by creating a catalog of all automations that clearly identifies the various attributes affecting the smooth operation of automations. The enterprise should also maintain a list of dependencies including internal and external applications, the APIs leveraged, network paths, databases, shared files, credentials, automation schedule, and so on.

Developing a framework that creates a heat map of the impact to various automations caused by changes in dependent systems helps to proactively test the automations as opposed to later having to react to the changes as the exceptions occur.


Beyond merely introducing you to RPA, this course insightfully delivers an overarching understanding of the Automation Anywhere
Enterprise® platform, including impactful key features and commands, empowering you to create an end-to-end automation process within your enterprise.

Upon completion, you will be prepared to sit for the exam and confidently complete the process to earn the Automation Anywhere
“Mastering Bots: Design and Build an Advanced Digital Workforce” Certification. View the Automation Anywhere course outline or contact us today to inquire about automation training.


Jan 2020

By: Terry Mott