Feedback on Robotic Process Automation

I sat with an old colleague to discuss some pieces of the RPA landscape that Mike thought were under-represented in the article.  It was great to catch up with an old friend, as well as to get a different perspective on the different tool sets in automation.  

Transcript

Welcome to this edition of Software Strategy Podcast. We are meeting today with an old colleague of mine, Mike Howard. Mike and I first worked together back in the early  1990s. We worked for a small technology consultancy called Business Technology Consultants. Mike was my mentor and source of all knowledge and wisdom in the world. Since then we have both worked for different companies on different continents and briefly for the same company on different continents we have both had wild and interesting careers and taken up strange sports. 

Mike is an expert in enterprise resource planning systems, IoT, Manufacturing and supply chain planning processes.

Mike.  Welcome to the Software Strategy Podcast

Thank you for having me

Mike gave me some very interesting feedback on my software strategy insight that I published for robotic process automation. Mike noted a couple of omissions in my interpretation and explanation of robotic process automation.  I am very grateful for the feedback, and it gave me an opportunity to invite Mike onto the podcast to explain the pieces that I had missed. 

Mike, you gave me the feedback that an area that I had missed in my explanation was really how artificial intelligence interacts with robotic process automation.  Maybe you could give our listeners the benefit of your insights into how these two areas of technology interact. 

Thanks Nigel, always good to catch up and learn from each other. RPA, in fact automation of any repetitive task has been around for a very long time. Printing presses to save handwriting books, to automated manufacturing to auto pilots in aircraft. What RPA does is remove manual repetitive tasks whilst using insights from AI and ML to gauge the validity of a decision. A simple example – a manager is approving expense claims. A staff member regularly claims for internet and their mobile phone plan. After 2-3 consistent approvals the system learns the parameters and is able to recognise what is OK to approve. And simply sending the manager an email saying – approved. Avoiding the need to actually look into the details of each claim. IF the person takes an overseas trip and the mobile charges rise then this is flagged. Therefor the manager only has to look into exceptions.

Another example for RPA and one that is used extensively is testing software. It is possible to test many different business processes extremely quickly prior to releasing it to the public. Saving a lot of time and cost. Using shopping cart example – smart systems can now automatically test a storefront with varying sale items, in and out of stock and currencies. And validate the pricing, delivery cost etc.

Thank you Mike. The other area that you had mentioned was the relationship between robotic process automation and key performance indicators.  Again if I could ask you to explain to our listeners the relationship between a process as defined within robotic process automation and how it is measured in key performance indicators. 

Absolutely Nigel, this is where I see there is so much potential and where new business roles will be generated. The advancements in compute and storage have enabled automated decision processes to become affordable to many more applications. In the business world it is now possible to compare tasks like financial month end close across similar companies, compare their KPIs and make informed decisions on process improvements. When applied to manufacturing, it is possible today to plan around supply issues. Where for example a common component is in short supply and will result in an undersupply of many different end items. Normally this would take a planner quite some time to resolve. Especially if there are contract penalties on some parent items. With the use of AI and ML many of the planning decisions can be processed providing the planner with optimal choices. I see this as just the beginning – as a passing thought consider this applied to self-driving cars. The potential is huge.

Mike, I would like to thank you for your insights on behalf of both the software strategy podcast audience and myself. I cannot believe that I am still learning from you some 30 years after we first met. 

That brings us to the end of this software strategy podcast thank you again for tuning in please do not forget to subscribe for further software strategy insights. Brought to you by a network of seasoned professionals in the field.