Fifteen digital leaders gathered in Oslo in January to discuss the topic of Intelligent Automation. The focus for the evening’s round table was how to build a practical roadmap to achieve successful business outcomes in this rapidly advancing area of digital transformation. Jan-Olav Styrvold led the discussion on behalf of Blue Prism and Avanade.

Setting the context

The word ‘Intelligent Automation’ covers a wide range of topics. In this case we focused the discussion on how it relates to Artificial Intelligence (AI).  AI is a term that is often banded about, but without a tight definition of what it means.  This has led to unnecessary reluctance, and sometimes even fear, to embrace the opportunities that lie in implementing intelligent automation.  There are several examples of successful AI solutions to be found online, but also of many failures.  However, the failures do not seem to stem from technical limitations but rather from the “soft” areas of organization and training.

In the Nordics, Finland seems to be forging ahead as early adopters of intelligent automation and related AI while Norway, Sweden and Denmark have not been as quick to adopt this type of solutions.

What changes can we expect in the organization?

A recurring theme in discussing intelligent automation, or Robotic Process Automation (RPA) as it often is called, is the effect on the human workforce when introducing digital methods. Prof. Leslie Wilcox has coined the phrase ‘taking the robot out of the human”, that is, letting robots take over the mind-numbing routine operations that do not require skills that humans are good at, e.g. empathy, creativity, lateral thinking.  When introducing RPA, we are not necessarily reducing staff numbers, but we must be prepared to work hard at making sure the organization has the necessary skill-sets that a digital work force requires.  Research has shown, however, that in some areas a reduction of human staff in the region of 80% can be expected.

How do we ensure a successful implementation?

In this area, as in all other major changes in an organization, support from top level management is key. Sometimes where there is push-back from the business it is important to make sure that the commercial reasons for introducing RPA are aligned with board objectives. There are now several successful user stories that should help the organization understand the value of RPA and how to implement intelligent automation.

Another key to success is making sure RPA really is business driven, it cannot be viewed as an isolated IT endeavor.

How do we identify the best tasks to automate?

A common mistake when introducing RPA is to analyze current business processes, identify the routine parts in them, and train the “robots” to handle them. A better way is to start by redesigning the processes with a digital workforce in mind. Thereafter, we need to employ service design and UX-experts to make sure that the new processes not only optimize robotic and human input, but also design interactions from the user/customer point of view.

A common question is: how large a part of the process should we automate? This will vary from process to process. However, it is better to start with a smaller percentage and have those steps done well, rather than aiming for a high degree of automation, and not do it well at all.  It is much easier to start from a success story and expand from there, rather than repairing a solution that is struggling.

How do we move from our “proof of concepts” (POCs) to a production solution?

Many organizations, both in the private and public sectors, have set up POCs. These are working well, but what are the next steps?

The most important thing is to plan for success when starting a POC. That means starting by implementing the organizational changes that are needed, making sure that resources are available with the right skills, and agreeing on KPIs with the business units involved.

The experience so far is that the RPA technology is mature, it is just a matter of making it work for the business.

The future of intelligent automation

Most of today’s RPA solutions are deterministic in nature, that is; rules based. The solutions are trained on the questions that humans have answered historically.  But that may not be the case for very long. Another branch of AI is making great strides; machine learning.  Now that we have robots that have taught themselves how to play poker and have beaten professional players, we can expect our RPA implementations to move in the same direction.

Perhaps the “point of singularity” that John von Neumann coined in the 50’s and more recently Professor Stephen Hawking and Elon Musk warned about is not very far off, when a machine surpasses human general intelligence. This will have profound implications for our organisations and society at large.

This article was written by Jan-Olav Styrvold, director at Analytika. With a long and varied career in the IT industry, including managing director of SAS Institute Norway, Vice President of SAS Institute EMEA and lately both CFO and CIO at the Norwegian Wine Monopoly. Furthermore, Jan-Olav worked with Fornebu Consulting for more than a decade, as a board director and board member of a number of companies and is also member of the Norwegian ICT Council.