Instrumented Automation

I have talked about the basic concepts of automation and introducing intelligence to rule based automation using data and artificial intelligence.

In this paper I will talk about instrumented automation.

To optimize the benefit of automation, humans will have to work with machines in way that is seamless, non-intrusive, non-distractive and self-improving. It needs to instrumented.

What does instrumented automation mean?

Automation will be used by man and machine – seamlessly with minimal intervention of IT support. It should be easy to use, flexible to change and most importantly pervasive. Instrumentation should ensure that automation is embedded as a part of the business process, and not a separate utility that needs to be executed.

Instrumentation enables adaption to changes in a process. Technology exists to discover process activities and mine those activities to identify changes. Technology also exist to change orchestration on the fly. All that is needed is to tie them together to the automated business process context.

Getting started with instrumented automation

 Let us start with the concept – how automation work today and then how instrumented automation improves upon it.

Today through discussion or using expert advice or analysing the current bottlenecks or a combination of any of these, automation use cases are identified. These use cases then go through the normal design-build-deploy life cycle. Several processes change frequently and so the change management process sets it.

With instrumented automation the process activities of users and machines are capture and anlyzed to identify repetitive tasks. As we know all repetitive tasks may or may not be candidates of automation, the next step is to derive the tasks which are automatable. These automatable tasks are then auto generated and deployed.

As the automated tasks are running, process activities are now observed for the hybrid workforce, to identify changes as well as process improvement needs, and they are again auto generated and deployed. The cycle continues with oversight for exception management.

There are a few challenges here, the primary being the privacy of humans as their activities get recorded. These needs to be complied with selective data collection, anonymization and other available techniques. 

Characteristics of instrumented automation

To conclude let's discuss what are the distinctive features of instrumented automation. 

• Self Service – Consumers whether humans or machines should be able to discover automata’s and use them within their business processes dynamically.
• Flexible – Changes to processes needs to e identified and addressed automatically with minimal intervention.
• Pervasive – Automation execution should omni-present and there is no need to build or run them manually or through static rule based orchestrations.
• Conversational – The usage should be conversation driven, with a natural  style of interaction like the ability to change topics during a discussion and suggest the next step to the users.