House of RPA 2.0 Or How a Stanford Professor Made Me Think About Wallace and Gromit

A couple of years ago, I wrote an article on here about how businesses need to think about RPA like buying a house.? Here’s the link.? (TLDR:? Data foundation is important, all of the business should be involved; maintenance and monitoring are important.) ? It’s an analogy that I’ve used quite often when describing RPA to people.? Now I feel like the analogy needs a little bit of an update.? If nothing else, the analogy needs to be updated for AI.? (All LinkedIn posts must reference AI.? I don’t make the rules.)? But first…

One of the highlights of last year was going to Forward 6, UiPath’s yearly conference.? At the conference, my favorite speaker was Erik Brynjolfsson, a Professor and Senior Fellow at Stanford.? (That is selling him so ridiculously short.? He’s a genius and you should be following him here:? https://www.dhirubhai.net/in/erikbrynjolfsson/.)? In his keynote address he compared the onset of AI to the invention of electricity.? When it was first invented, factories replaced their centralized steam turbines with centralized electric engines.? It worked great. Factories were still a maze of belts but now with less steam and water. Progress! Plus the processes and factory layouts really didn't didn’t have to change. However there wasn't a ton of productivity gain. It wasn’t until the idea of the assembly line, placing a small engine at each workstation did productivity start to really increase.? Electricity allowed for more efficient processes.? (Here’s a 2017 BBC article about it.? https://www.bbc.com/news/business-40673694? The fact it references Erik is a happy coincidence.)?

Think about houses. It took a generation for houses to get all the necessary wiring to make home living more efficient with electricity. Great advances take time to implement. Erik’s speech is one I’ve referenced multiple times over the past months.? Which leads me to Wallace and Gromit.??

Here’s a clip from Wallace and Gromit’s cartoon “The Wrong Trousers”.? https://www.youtube.com/watch?v=EGSyw2dHhrc. In the cartoon, Wallace hits a button right next to his bed for Breakfast and an automation raises his bed, opens a hole in his floor, and he slides out of bed, into his pants and falls right into his chair downstairs.? Once he lands, his loyal dog Gromit hits a button that shoots jam that lands on a piece of toast that falls on his plate.? Breakfast is served.? To say it’s pretty slick is an understatement. All Wallace does is hit a button and quickly he’s dressed, moved downstairs to his seat at the table with breakfast right in front of him.? That’s automation!? It took some work though.? The bed needed to be updated so it could raise to slide him out of bed. A hole in the floor had to be made. Was his bedroom always right above the kitchen? Maybe some redecorating had to occur. Finally it takes standardization to make the processes work.? Wallace is having toast and jam every morning for breakfast. In time, maybe he adds another breakfast option.? However in the beginning, it’s toast and jam.??

I’m sure if there was a “The Wrong Trousers” prequel it would be pretty boring.? It would also probably show a very remedial automation.? Maybe Wallace had a button at the table that served him breakfast when he was seated.? He still had to manually get out of bed, get dressed and go downstairs.? Ugh, what work! ? However, he thought big. What if the automation got me out of bed, got me dressed, got me downstairs right to the table and delivered my breakfast right when I was ready?

I feel like automation and AI are in the prequel stage right now. Many businesses just swapped out the person for a robot doing the same tasks. Are these processes productive? Sure but there's so much more that can be done. Many processes haven’t been changed or even invented to fully utilize AI and robots.? So much of the work of automation and AI over the next few years is going to be automating processes that have never existed. These are always my favorite types of processes to work on. For example, why is month end processing a thing? With automation, that processing should be occurring daily/hourly etc with no end of the month rush. The AI/RPA assembly line still needs to be created. It takes time to think about all of the things an AI powered robot can do.? It takes creativity to think about what is possible with AI.? (There’s a ton of AI “thoughts” out here on LinkedIn.? I’d love to see more “process creativity” thinking.)? It also takes process owners willing to make some changes (“You want to cut a hole in the floor why?”) to get the long-term benefits that automation and AI can provide.??

For automated processes what are the changes that process owners will need to make? A big piece is data validation. (I saw a post that said everyone is going to be Quality Assurance in the future. That is totally spot on.) Instead of manually entering data maybe that data needs to be validated before it's entered. The other piece that may need to be done is validation after the automation completes its work. Automation is going to completely do everything, a person is still going to have a role in the work. I feel like a lot of early RPA work has been to have a robot completely take over a process. That's unrealistic. People will still have a role. For Wallace, he had to hit the button. Process owners will need to be involved. Validate the data before, check the reporting after. Plus, there's always going to be an exception to the process that will likely need to be worked. ?

A couple of years ago, I wrote about how all facets of a business can and should be involved in a company’s automation journey.? That’s absolutely still true.? Now everyone needs to be creative and think about what AI and automation can do. Don't be limited to just what a person is doing now. Think about the input and output of a process then think big about what else a robot could do.??

P. S.:? I see so many posts on LinkedIn talking about how it's hard to tell if this was written by a person or AI. What prompt would come up with this?