Note: You are looking at a static snapshot of documentation related to Robot Framework automations. The most recent documentation is at https://robocorp.com/docs

Technical feasibility assessment

Maria wants to automate her task of copy-pasting data into the intranet, but how can she know if the automation is feasible in the first place? Let's find out!

The technical feasibility assessment is done to understand if the automation goal is realistic. It is good to find any technical showstoppers early on. In the spirit of agile development, fail fast (and save money doing so)!

Although the assessment is technical, consider the available budget. You might find showstoppers if your budget is very tight. Even if something is technically possible, it might be too costly.

Think of the following questions to figure out if the process could be technically feasible to automate:

Could a robot complete the process by following unambiguous rules?

  • If the business rules are easy to follow, unambiguous, and do not require human judgment, great!

  • If the rules are complex but unambiguous, more effort is necessary during the implementation, but the automation could still be doable.

  • If human judgment is required to complete the process, it might be difficult or costly to automate the process. In those cases, you could think if parts of the process could still be automated.

In Maria's case, the rules to follow are simple and unambiguous. Download the sales data Excel file, open the intranet website, log in, enter the sales data, take a screenshot, create a PDF document. Definitely something a robot could do!

Can a robot easily read the data required by the process?

  • If the input data required by the process is in a digitized format that the robot can read comfortably, wonderful!

  • If the data requires some transformation before the robot can use it, things get more complicated. Still a good chance for success!

  • If the data needs to be extracted from images, PDF documents, or emails, there are solutions for those, but prepare for a bit more intensive implementation. No blockers, still (unless the budget is exceeded)!

  • If the data is not in a digitized format (it's on paper, for example), you need to get that data digitized. Depending on your budget, this might be a showstopper.

In Maria's case, the process input data (the Excel file) is something that a robot can read easily. No data transformations are needed; grab and use the data as is! The robot can take screenshots of the page (or parts of the page). The robot can also get the HTML from the intranet webpage for generating the PDF file.

Can a robot operate all the required software?

  • If there are only browser-based web applications involved, this is usually the best case (plenty of support for automation). Robocorp supports Selenium and Playwright out-of-the-box!

  • If the process involves automating desktop applications or even old-fashioned mainframes, the implementation might take more effort (depending on the operating system and the applications). Robocorp supports image template matching based automation, textual locator-based desktop automation, optical Character Recognition (OCR), and keyboard and mouse-based automation!

Maria's process involves using several technologies. Excel (file) for the sales data, a web browser for entering the data to the intranet, Word for pasting and exporting the results in PDF format. The robot can read Excel files without actually using the Excel application. The robot can use whatever browser is installed. It can also produce PDF documents without the need for an additional application. It seems that there are no technical issues with this question either!

Can the robot access all the required accounts?

  • If the robot can use its own credentials on all systems to complete the process, great!

  • If the authentication can not be automated completely, some human intervention is likely to be needed. This will have an impact on development.

In Maria's case, the robot can use her credentials, stored somewhere safe, such as in the vault in Control Room. Good to go!

Proceed or abort?

After thinking through the questions, you will have a better idea of whether the automation could be technically feasible.

Sometimes the result of the technical assessment is that the automation is not technically feasible at all. This is both an unfortunate and a good thing! The sad part is that you still need a human to complete the mundane process. The good part is that you found out the showstoppers as early as possible, without wasting time and money!

Time to move to assess the return on investment, then!