• Artificial Intelligence
  • Setrag Khoshafian
  • FEB 11, 2020

Spectrums of Work Automation: The Future is here …

There are many predictions that attempt to capture how many jobs will be lost due to digitization and automation. As an example, a recent study by McKinsey indicated automation threatens 800 million jobs and a 1/3rd of workers in the US and Germany need to learn new skills by 2030. There are similar studies from the World Economic Forum – also emphasizing reskilling - and others

These are interesting. There are helpful observations, cautions as well as recommendations to be prepared in different industries as well as regions. But often these types of surveys and predictions are naïve and simplistic in the characterization of work as well as categories of workers: be they white or blue collar. The spectrum of digital transformation as well as cultural transformations are often ignored. 

Robots and Robotic Automation

In various industries or even in our homes or offices, robots come in all shapes and sizes, from heavy robotic arms on a shop floor to greeting robots in retail stores, restaurants, or hotels. 

Robots are also appearing through systems of connected and increasingly intelligent devices. For instance a self-driving connected vehicle is a robot: but it involves multiple smaller robots or intelligent connect sensors and actuators. This aggregate of devices, smaller robots, and intelligent sensors need to all orchestrate and collaborate to achieve the objectives of the larger aggregate robot: the connected vehicle. In fact we can extrapolate from this to other even larger robots: such as a manufacturing shop floor, an intelligent building, or a connected city.

More interestingly, we can also have software robots. Repetitive tasks on, say, desktops can be automated through robotic software – saving considerable amount of time and avoiding potential human errors. Thus the era of robotics ­– be they physical robots or robotic automation in software is making a huge impact on process efficiency. Thus instead of humans copying information between various screens, or traversing a plethora of legacy applications a software robot instead automatically achieves the same repetitive tasks: much faster and with fewer errors.    

This brings us to the crux of Work automation – viewing it as a spectrum of work types and a spectrum of worker categories.

Work Types and Worker Categories

When it comes to Automation, there is a spectrum of work and worker categories.

  • Repetitive Work Automated through Robots and Robotic Automation: increasingly robots are replacing routine predictable and repetitive manual work. This is sometimes characterized as the “long tail” of tasks that are allocated to relatively low paying repetitive jobs that are managed by the business vs. IT. These can now be automated with robots. These include actual physical robots that can, for instance, manage a warehouse moving boxes around or clean the floor in large plants and of course you have extensive robotics in Manufacturing. Smart Manufacturing (aka Industrie 4.0) especially due to Industrial Internet of Things (IIoT) has given rise to the Adaptive Digital Factory. That is the industrial Operational Technology (OT) side. On the Information Technology (IT) side we are seeing increasingly robotic process automation solutions leveraged successfully in customer relationship as well as back-office operations. It is truly a pragmatic approach for automating work – especially in organizations that have many legacy systems.

  • AI Assisted Work: Now for more mission critical and higher business value (and typically higher transaction volume) tasks, you have AI-assisted workers that are guided by increasingly intelligent software that leverages business rules, analytics, and machine learning. It is similar to having an intelligent Siri or Alexa helping the worker complete their tasks. Increasingly workers are assisted by bot and Intelligent Virtual Assistants with increasingly sophisticated Natural Language Processing capabilities combined with knowledge that is contextual for the task or interaction at hand. The intelligent system now provides the contextual Next-Best-Action (NBA) to the AI assisted workers: prioritizing various decisions that could apply in a context. In many industries the connected devices themselves are also assisting the worker, even pro-activity: for instance a smart vehicle warning or making recommendations to the driver. AI assisted solutions are becoming increasingly intelligent: especially in the context of end-to-end value streams empowered with underlying rich intelligent capabilities such as rules, predictive analytics, machine learning adaptive analytics, and Big Data recommendations. The overall end-to-end value stream of different types of works orchestration with humans, robots, virtual assistants, and robotic automation when applicable.

  • Involving Cognitive or Knowledge Workers: Knowledge workers are the experts. They are the cognitive workers. Some have accumulated considerable knowledge and expertise in particular domains. Some organizations are facing a crisis as these workers start to retire and their sometimes irreplaceable know-how knowledge is lost. A complete digitization strategy needs to also capture, digitize, and automate this knowledge. It is another category of intelligence – this time emanating from human intelligence or knowledge. Work is not always predetermined or structured. There are sometimes – more often than we give credit to – unanticipated exceptions: ad-hoc, dynamic, or unplanned tasks. The cognitive workers need to be involved to solve the difficult challenges. In some cases their expertise is invaluable. No “Artificial” intelligence will be able to replace this category of workers. They innovate and often come up with the policies and procedures in the organization. They can react on the spot, knowing what to do in a particular, exceptional situation. Knowledge workers can be engaged to resolve cases or processes. They can also analyze the reports and the performance of operationalized processes. The knowledge and know-how of this important category of workers need to be harvested and digitized. Ideally, they need to work closely with data scientists: to complement human knowledge with data and machine discovered models. Most importantly their knowledge need to be harvested and digitized in the context of end-to-end value streams.

As indicated above, this is a spectrum - there are many hybrids and ranges of automation in between these categories. So this whole idea of humans loosing jobs to automation - physical robots or robotic automation in software - needs to be approach much holistically. The human factor is very much in play and repetitive jobs will be replaced with much more exciting ones. I’ve covered these in a recent webinar on Bots, AI, Robotics and the world of Work

The Harvard Innovation Lab

Made in Boston @

The Harvard Innovation Lab


Matching Providers

Matching providers 2
comments powered by Disqus.