Robotic Process Automation: The 4 Critical Stages of Implementation

You are here

Robotic Process Automation: The 4 Critical Stages of Implementation

Posted September 28, 2017 in Business Technology & Digital Transformation Strategies
Figure 1 — The four critical junctures of RPA implementation.

Robotic process automation (RPA) implementation is a process with defined steps. In this Advisor, we discuss the critical junctures that are required to achieve process automation implementation in any organization. The industry currently follows an Agile methodology to implement RPA, with some customization depending on business needs. Here we describe the critical stages that are recommended for every RPA implementation.

The pre-sales roadmap for any RPA project during a project initiation (RFP, etc.) usually includes the following:

  1. Initial value proposition

  2. Due diligence

  3. Solution design

  4. Contracting

  5. Implementation

Below I explore the critical steps with respect to the development of a bot.

Critical Juncture 1: Tool Selection, Proof of Concept, and Cost/Benefit Analysis

Tool selection is part of due diligence. It is the most critical stage, as the benefits associated with RPA are directly associated with how fine automation is to be done. Any tool that is a good fit for the project will be able to create good scripts. Organizations should explore process feasibility and financial modeling so they can define the ROI and, based on the results, complete their tool selection.

Commercial tools for RPA in the market today include Automation Anywhere and BluePrism, among others. There are are also free/community versions, such as UiPath and Work Fusion, for organizations with limited budgets or process automation needs.

There may also be cases that do not require any of these tools at all, and instead process automation can be achieved by some of the widely used scripting languages, including Visual Basic macros or Python. However, these cases are very unusual, as these languages require additional scripting efforts compared to traditional RPA tools, and therefore may not be the preferred choice for most organizations.

A proof of concept (POC) is required for organizations to determine whether processes are feasible technically. In this stage, organizations should carefully identify a process for the POC, considering all the different technologies and controls (both Web and desktop), as well as a detailed review of the feasibility of current processes.

After establishing the POC, organizations should perform a cost/benefit analysis to determine the potential benefits of automation implementation. This analysis should be clearly quantified, as the break-even on ROI depends upon an accurate cost/benefit analysis.

Some of the factors that need to be considered during the cost/benefit analysis include:

  1. Duration of the processes required. In some cases, the process lifetime is short, with fewer resources (employees) working on it. These are not the ideal candidates for automation. Even if they are feasible for automation, commercial tools are not preferred for such cases.

  2. Trends need to be carefully analyzed. Processes that highly depend on government regulatory guidelines are not good candidates for automation as there is a chance of high maintenance requirements.

As per RPA industry best practices, savings for different processes are generally as follows:

  1. Standalone applications (RPA replaces one bot for every five users)

  2. Web applications (one bot for every four and a half users)

  3. Citrix applications (one bot for every three and three-quarter users)

  4. Remote (one bot for every three and a half users)

Critical Juncture 2: Process Selection and Risk Identification

Process selection is another subset of due diligence. The process selection critical juncture is the stage for determining automation candidates. Those processes that are preferred candidates for automation meet the following requirements:

  1. Stable processes

  2. Repetitive in nature

  3. Structured in nature

Processes that are not preferred for automation include:

  1. Unstable processes (UI controls are changing every now and then)

  2. Not structured, or indefinite flow

  3. Processes that depend on government guidelines and have high change rates

The success of RPA depends on the selection of those processes that are to be developed and deployed. Organizations must understand that 100% automation is impossible. Therefore, which processes are selected and which are rejected is a deciding factor for RPA success.

Critical Juncture 3: Detailed Planning, Bot Creation, and the Dry Run

Once your organization has selected a tool and identified processes, you should identify your process steps, which require detailed analysis. In this modeling, the organization should create a thorough plan that includes exact costs. As in the previous stages, risks are identified in order to mitigate those in the development effort and rising costs. This plan can illustrate exact breakeven durations and ROI values, including:

  1. Cost incurred as an input while creating automated flow

  • Cost involve in purchasing RPA tool license

  • Cost of resources required to create automated flow

  • Software/hardware required for RPA setup

  1. Cost saved during the execution of automated flow. Whether a breakeven of ROI is achieved depends entirely on the process complexity and its execution time, therefore differing from organization to organization.

After the bot is created, it also needs to be tested for its consistency and accuracy. Organizations should plan and initiate several rounds of attended and unattended runs, which will define the stability and readiness for bot deployment.

Critical Juncture 4: Execution and Maintenance

In this final and stable critical juncture, a bot is deployed for execution. If the processes are dynamic, changes need to be maintained during this phase.

Figure 1 — The four critical junctures of RPA implementation.

Figure 1 — The four critical junctures of RPA implementation.


Process automation, if implemented wisely, can provide huge cost savings as well as make your business more effective and efficient. We are seeing continuous improvement in automation, and the day is not far off when we will have process automation enabled with artificial intelligence as well as automated machine learning. When that time comes, we will have minimal to no human intervention during the manual process. To reach that critical juncture successfully, we first need to bring out the best in existing available tools and opportunities using the approach outlined in this Advisor.

Across the industry, robotic process automation is in its elementary phase, and is slowly making its way toward business transformation. In the future, we may witness the evolvement of robotic process automation into smart process automation.

20 Keys to RPA Success

20 Keys to RPA Success — Save 50% when you order today!

How are today's organizations successfully adopting robotic process automation and cognitive automation? What are some of the outcomes they are achieving? What practices are they employing? Find out in this new report. Order your copy now using Coupon Code RPA50 and SAVE 50%!

About The Author

Prashant Chaudhary's picture

Prashant Chaudhary has rich experience in automation, including framework assessment and tool evaluation. He has 10 years of experience working with various automation tools, including IBM RFT, IBM RPT, HP UFT, Selenium, Automation Anywhere, UiPath, and others.

Mr. Chaudhary lives in New Delhi, India, with his daughter, wife, and parents. He is an active blogger on the topic of the various aspects of automation. His domain expertise... Read More