I recently worked with a group including Vishwas Lele and Jesse FitzGibbon to design and execute an accelerated learning series on Robotic Process Automation (RPA) with Microsoft Power Automate, kindly dubbed the “AIS RPA Crash Course.” We set out to engage a team of roughly 20 AIS employees with a diverse set of skills and experience levels with the goal of providing them with the practical experience to contribute to enterprise-level automation projects with AIS customers. It proved to be a rewarding experience for all of us. So, we thought it would be helpful to share some details of our approach and what we learned along the way.
The first week was all about establishing a common understanding of RPA and its importance for our customers and employees. RPA continues to be popular as the importance of hyperautomation is being felt by IT and business users alike. By infusing RPA with AI /ML (Cognitive Services) and a low / no-code platform, Power Automate is democratizing access to RPA (please refer to the latest Gartner Magic Quadrant for RPA).
We began with Microsoft Power Automate and its role in hyperautomation – including the synergy between cloud and desktop flows. After our synchronous sessions were completed, we spent another day working in a hybrid model, loosely following Microsoft’s “RPA in a Day” structure which combined guided instruction with independent, hands-on lab work. This really helped to solidify the concepts we’d spent the previous day discussing and put our team in a good position to start thinking through how these technologies could be applied to solve real-world problems.
We wrapped up the first week with a “Proof of Learning” exercise where we asked each of our participants to identify a potential real-world application for RPA and to build a proof of concept to demonstrate hyperautomation in action. This exercise really forced everyone to think through how to get the building blocks in place. It was awesome to see the excitement, interest, and confidence that everyone had acquired as they presented the results of their work.
During the second week, we provided the participants with a new business problem to solve. We asked the team to use RPA to react when a file containing GPS coordinate data is uploaded to Azure cloud storage, and then upload that file to a website (which we provided) that summarizes that data in terms of total distance, elevation, average speed, etc. Once the website had completed its magic, the process should then capture the results and email them to the submitter of the original file.
The process was somewhat basic, but we asked everyone to think beyond the fundamentals of “how do we automate this” to “how do we implement this in a robust and repeatable way”. We also required that the automation be implemented using Power Automate Desktop and assume REST APIs were unavailable.
We divided the team into pods. Each pod was assigned a project manager and was comprised of people working across US and India time zones. The pods were tasked with developing their own feature backlogs and independently attacking solution design and implementation. Vishwas, Jesse, and I closely monitored each of the team’s progress during the week and provided interventions and support as needed, most often in the form of brainstorming and assistance with specific technical obstacles.
The group presentations at the end of the week were remarkable. Each team had fully automated the process and had begun implementing more advanced concepts. What I found most interesting was how each of the teams’ diverse backgrounds we re represented in their solutions with an incredible emphasis on user experience. One team even built out a canvas app as a front-end to upload the .gpx files.
Our third and final week focused on the implementation of best practices for application lifecycle management, securing secrets with Azure Key Vault, and ensuring the security of the data from start to finish – all of which we believe to be critical to making RPA solutions ready for the enterprise. The teams continued to iterate on their solutions from the previous week and incorporated new components intended to make the applications more robust and manageable. The built-in error handling leveraged the concepts of chaos testing to ensure that failures were handled elegantly and that logging was in place to allow for future troubleshooting. Other enhancements were made to plan for how their solutions would automatically and efficiently move from test to production tiers using Git and GitHub Power Platform Actions.
Results and Takeaways
We are pretty proud of the results. Our participants started with very little if any, knowledge or experience with RPA and Power Platform. At the end of 3 weeks, we had a team of people with a fresh set of skills that they were excited to use – and it was not just technical skills. Throughout this exercise, our teammates embraced our continued learning core values by striving to improve themselves to better serve our clients in a rapidly changing technological landscape.
Many of our participants have moved on to new projects since completing the course. Better yet, several of them have joined forces with other AIS engineers working to solve critical client needs using RPA!
This experience was equally enriching and energizing for those of us charged with structuring and overseeing the course. We look forward to leveraging our lessons learned to improve on the next iteration and thank all our participants for their hard work and willingness to roll with the punches as we crashed our way through the course together.