Digital Process Automation and Human Activity
- December 14, 2017
- 1 Comments
I had a discussion today with one of our customers who has a very interesting approach to their Digital Process Automation practice. While not explicitly stated by them, I think a formal statement of the Thesis would be:
The goal of every process solution is to either have an action taken, or have a decision made. This means that every human activity in a process can be summarized as one of the following 3 actions –
- To make a decision based on the data presented.
- To take otherwise opaque data and provide it to the the DPA System.
- To take an action that requires a human to execute.
At first, I very much objected to their thesis. However, the longer I thought about it, the more it seemed to resonate. This has interesting repercussions for our delivery teams today, as well as for understanding what processes may look like as automation become easier and more obtainable.
Impact on Process Automation Delivery
The customer I was talking to seems to have a very efficient and well run Process Automation program. I have come to believe that part of this is due to their adherence to this principle. If you agree with the principle, then for every human activity you are placing in your Process, you should be able to answer “What is the goal of this task?” The important thing here is that the answer really should only be 1 of the above action – make a decision, add data, take an action. It is when these items become intertwined in a single activity that users are presented with overly complex interfaces that make accomplishing their job harder.
Future of Digital Process Automation
The very interesting thing is that, looking at each of the above items, we can see where DPA is going. Today, this customer already has a goal of moving most of the decisions out of the realm of experts and into a decision management system. Where humans are still making decisions, the use of DPA means they are gathering the data that can enable future Machine Learning to automate these decisions.
For “opaque” data, the solution will depend on why the system doesn’t have the information. If the user is effectively the “integration bus” for the system, meaning the screen is asking them to re-key data from one system to another, then this is ripe for replacement by RPA in the near future. If however it is a higher order action, for example conducting and creating a site survey for an oil well, there isn’t a near term technology solution for that problem.
The final item is likely one of varying opportunities for technological replacement. There are things to that seem like they are likely to be automated soon e.g. – driving delivery trucks. There are other that seem much harder to automate – figuring out where at to leave the package if no one is available.