Before You Go — RPA Post-Implementation Analysis

Posted August 10, 2017 in Business Technology & Digital Transformation Strategies
Prashant Chaudhary

Analysis is, in general, a critical part of every project. Most commonly, however, that analysis is conducted at the beginning of a project. In this article, we explore the benefits of the post-implementation analysis of bots/robotics process automation (RPA) flow.

The best time to conduct a post-implementation analysis is after a few rounds of execution, when there is data available and you can start analyzing those statistics —  ideally after seven to 10 days. With smaller-scale processes, then you may wait a bit longer. Here, we consider a mid-sized process in terms of complexity.

Analyzing an RPA implementation requires an assessment of the economy of scale and size (ESS) of operational processes. This assessment should explore factors such as the size and complexity of the project, the total effort required to manually process it, and the effort required to automate it if it is feasible for automation. Identifying and selecting a tool for your process automation depends a lot on this ESS data, as you may be able to avoid commercial tools and opt for a community version. Tool cost after implementation may be even higher, and may prevent breaking even on the project.

Teams make assumptions during the initial ROI and cost benefit analysis. By performing a comprehensive analysis of existing RPA implementations, an organization can achieve additional savings in future bot deployments.

Ultimately, RPA implementation reduces the size of the workforce that is required and saves costs according to the industry-average ratios shown below (figures can vary with the size of processes):

  1. 1:5 in workforce reduction — in a process that requires 100 human resources, the workforce can be reduced by up to 20.

  2. 1:3 in cost reduction — an expenditure of US $100,000 can be reduced by up to $30,000.

Post-implementation analysis thus potentially achieves savings of up to 15%-20% of the initial ROI calculations. (Since savings can vary implementation to implementation, it is not possible to provide exact amounts.)


Post-implementation analysis is sometimes referred to as “assessment” in IT/ITES automation projects. Whereas an assessment is generally an evaluation of current process, deliverables, and so on, RPA bots are widely used for clerical process automation and do not usually involve direct client delivery. Across industries, much of RPA bot development uses Agile methods. Agile is responsive and swift, and provides projects with an increased prospect for advancement when analysis is performed post deployment.

Some of the processes where RPA is widely used and preferred include:

  1. Sales and invoice processing

  2. Inventory management

  3. HR administration

  4. Claims processing

  5. IT processes

At the end of seven to eight days post-implementation, organizations should perform a thorough analysis of dashboard data. Table 1 illustrates this analysis and its associated benefits.



How many times a bot fails to perform a task


Bot accuracy will increase after analysis of the failure and its root cause identification.

In future implementations, the same issues/gaps can be fixed, resulting in more accurate bots.

How long a bot takes to execute

Execution time analysis can help in calculating actual ROI as well as providing the possibility for exploring further optimization.

Manual verification of bot activities in business flow


Manual verification of process again helps in validating the accuracy of bots, building more trust in a bot’s execution.

Finding the deviation in ROI by calculating actual data and comparing with earlier data (assumptions)


Learning the deviation in pre- and post-implementation ROI calculations can help operation units make more accurate assumptions as well as help in making decisions that are based on ROI, resourcing, and so on.

Machine utilization-bot runner machine

In some cases, bot execution increases machine CPU utilization, sometimes leading to tools crashing and causing automation exceptions due to increased logical statements. These development issues and/or system dependencies should be identified for those processes that involve huge data processing.


Table 1 — A post-implementation analysis of dashboard data.


Based on the results of the analysis described above, organizations can determine whether their processes require further optimization or script corrections. They may also explore the option for further automation within their automated processes as a result of this information.

There are processes where manual intervention is required even after automation, which we call "attended automation." Post-implementation analysis allows organizations to explore the options for making those processes unattended by exploring areas where automation has been ruled out initially. Such decisions can be made after post-implementation analysis, enabling RPA leaders and strategists to make more effective decisions in future implementations and help to enhance cost benefits.

About The Author
Prashant Chaudhary
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 is a… Read More