Fragile Automation

Have you ever experienced the situation were to satisfy an automated system you have to jump through several hoops when everything appears to have been much simpler before, maybe up to the point that processes are going to be established were providing of the data to the automation is done by dedicated people?

This is what I call “fragile automation” in the context of this text.

Especially in the context of striving towards increased efficiency organisations are trying to implement higher degrees of automation which also increases the risk of observing more of the above.

In the cases I have observed such scenarios were often the result of automation being done by one part of the organisation to improve its efficiency with a limited view on neighbouring organisations (thus having requestors jump through hoops in order to provide all inputs that make the execution convenient) and/or automation that is done to cover 90% of cases without the necessary precautions to be able to also deal with the remaining 10% in a reasonably efficient manner (which might include manual support as part of a proper fallout process).

Starting from the above observations (and I would love to hear/read your opinion on that) this means that in such scenarios the positive effects of automation get diluted; potentially even reversed.

This leads to the obvious question what can be done to avoid such outcomes. Some ideas are:

  1. Making sure the needs for efficiency of the automated processes customers are considered properly.
  2. Doing an assessment on the automated process including the consumers of the process to ensure avoidance of local optimization at the cost of others
  3. Making sure that the automated process is succeeding in (almost) all cases - as seen by the customer which may be via exceptions being handled manually (fallout process).

The main risk (especially with approaches 1 and 2) is an "overengineering" of the assessment and the solution, requiring good leadership to avoid the pitfalls. The 3rd approach is very much "good practice" which, in combination with proper measurement, opens the door to further improvement of the automated processes.

Pete Wallace

Business Transformation & Analysis, Focus on Operations, Services & Training

1 年

Interesting read Ronald Hasenberger I think there might be a prior point in the scheme of things, and that is to ensure the organisation really understands the processes they wish to automate, and then work to make the processes fit for automation. I have seen the "fragile automation" effect when existing process are automated as they are without sufficient analysis and re-engineering to make the fit for purpose.

As an add on to my article I would like to also point you to Sriram's article (https://www.dhirubhai.net/pulse/how-ensure-sustained-digital-organization-sriram-narayanan/?trackingId=px46sSyBSC%2BguNb0myfH4Q%3D%3D) providing some other means to avoid the scenario, looking at the topic from a different angle.

Sriram Narayanan

Principal - Bell Labs Consulting | Digital Transformation | Digital Operations | Business & Operations Transformation | PMP?, CPP Champion, Nokia Mentor, Lean Six Sigma | Views are my own

1 年

Nice one Ronald! an e.g. of third scenario is a Intent based workflow. So there is a trigger for a workflow automation, but further AI enabled intent based on exceptions that can handle the fallouts.

要查看或添加评论,请登录

Ronald Hasenberger的更多文章

  • AI-What AI?

    AI-What AI?

    Especially since the launch of ChatGPT the discussion about Artificial Intelligence (AI) quickly gained speed and the…

  • Automation and Control Theory

    Automation and Control Theory

    This blog post provides an introduction to closed loop operations for telecom networks, inspired by control loop…

    2 条评论
  • How sustainable is ICT?

    How sustainable is ICT?

    I recently had the pleasure to be invited to hold a talk on the question if telecoms (or ICT) and sustainability are…

    2 条评论
  • Facilitating cooperation with Metrics

    Facilitating cooperation with Metrics

    A frequently observed issue between organizations is local optimization where employees of each respective organization…

  • Dynamic Discrete Event Modelling for Operations

    Dynamic Discrete Event Modelling for Operations

    We are currently working on developing a dynamic discrete event model for operations processes. This is intended to…

社区洞察

其他会员也浏览了