I worked through a different use case recently and wanted to share out my thoughts and approach to it for others to leverage in the future. This one is unique because I did not focus on improvement but rather on putting intelligent automation around process. Like all other things, there are multiple ways to approach this and probably more elegant ways, but this sheds some light on how we can solve the issue of taking existing data from a cloud source and pushing it to an on premise machine to be processed. So let’s take a look at what I did and how I approached it.
Scenario
I need a way to submit my monthly project risks and have them validated in order to be uploaded into a master system. I have a predefined Excel document to be uploaded and the current process requires me to email it to our admin for validation. Now, as an admin, this is manually done countless times at the end of each month and want to put some automation around it.
For those thinking about using Nintex Forms or any other front-end, this would be a viable solution, however, there are times where we do not need to. Think about existing processes where you may be leveraging Excel files that have 50+ columns that have their own drop-downs and rules within. We do not want to rebuild that interface, at least right now. That is a separate effort entirely. While we would agree that there is a better way to (should) approach this, for now, we simply need a way to off load the burden of checking and validating the data within the file to something more hands-off.
The Process
Ultimately we want to have the Excel file checked for any issues. We need to setup a botflow that we can train to look for various business logic exceptions and then format the document to send back to the user. But what about getting the file to the RPA bot? Typically users would email the Excel file to the admin, but if we want to eliminate, or reduce, the interaction needed from the admin, we need a way to get the Excel file from the user to the RPA bot. Also, we will need to consider how do we manage letting the user know that there is re-work required. This is where Nintex Workflow Cloud comes in to focus.
We can easily build out a simple form to collect the Excel document and route that document to a cloud based file storage location. We will use OneDrive for this scenario. Once the the form has been submitted with the uploaded Excel document and stored into OneDrive, the RPA bot can easily get to the file. But how do we kick off the RPA bot? How does it know that it has something to go and do? Why we tell it to do something! Through the Nintex Gateway we can kick off a RPA job and process whatever we define within the botflow, which, the first step would be to grab the Excel file!
We pass data through the Nintex Gateway in order to provide necessary details for the botflow to execute properly. These details are:
- Excel File Location
- Submission Date
- Submitter’s Email Address
From here, the botflow has all that it needs to execute the process and validate the file. In my example, I am looking for special characters; specifically semicolons (“;”) and slashes (“/”) as well as validating that the data provided matches what is expected. I am looking at the “Inherent Risk Rating” for each project and have a list of expected choices. If the provided data does not match, I need to mark it along with the special characters.
I made a sample file with only 6 columns and 5 records for simplicity, but you could imagine just how time consuming this would be to manually check with hundreds or records and 50+ columns. Here is how it looks in real time:
Final Thoughts
I wanted to share this experience and showcase how easy it is to go from Nintex Workflow Cloud to Nintex RPA using the Nintex Gateway, but also to highlight something we often get stuck in; process improvement. I admit, the first time working on this solution, I wanted to build out a polished form in Nintex Workflow Cloud, but that would have taken a long time. I quickly realized that I can take what was already there in front of me and simply wrap automation around that. Improving upon this further will make us take a closer look at the Excel document and see if we can make a form out of it by removing data points or perhaps moving them to other steps in the process.
Jesse McHargue 