Fly by Wire or Fire and Forget – Augmented Versus Artificial Intelligent Automation?

This is the third and most likely final in a series of articles on the evolution of my Augmentation Intelligent Automation Theory and Practice over the last eight years. The first in the series, “Cybernetic Robotics – The Future of the Claim Processing Professional” was published in October of 2013. It was illustrated by the Giant Robot and Woman that became my Pareto Automation brand logo. It is the quintessential depiction of Intelligent Augmentation on several levels. Intelligent Augmentation is where humans supply intelligence to software robots. I am not talking about RDA. Where humans interact with desktop automation software to provide exceptions handling real time. Rather, very sophisticated RPA Digital Workers, which are detailed in the second and my most read article, “Autonomics or Cybernetic RPA?” published in July of 2016:

The two Visios that illustrated that article are now on either side of the Pareto Automation brand logo. They represent the technology of Autonomics (original definition) or Cybernetic Intelligent Augmentation Automation: a very advanced RPA Digital Worker to the left and equally advanced Rules Engine controller on the right. When combined, they provide a de facto remote-control process automation capability or as I will illustrate by analogy a very advanced “Fly by Wire” technology.

I waited a few weeks after publishing the pinnacle of my research and writing efforts: “ROC – How to Artifact” to see if I had anything further to say? This 55-page document (a free 31-page version is available at my website is the result of reviewing all my writings and related documentation over the last few of years to prepare for my speaking engagement at SSON’s 2018 Intelligent Automation World Series last month.

Last week, I woke up for the third time and realized, I had one more thing to say. 

I am convinced, that a Billion dollars of annual cost savings still resides in the five percent of Managed Health Care claim automation between 88 to 92 and 93 and 97% adjudication range. There is a four % difference in the start and end points of the five-point range depending on the type of claim system and contract complexity. Now that AI Hype is clearing (I was affected too) and the root causes of significant process automation implementation failures are increasingly traced to business model/culture change management issues. Human Centric Augmentation approaches to Process Automation are gaining the increased credibility they deserve. I am confident that Augmentation Intelligent Automation techniques can automate that 5% for perhaps a third to one half the cost of Artificial Intelligent Automation techniques.

Put simply, Rules Engine Controlled RPA Digital Workers with good OCR and perhaps NLP to access unstructured data, using an Ops Dev business model facilitated by a Culture of Trust that fully engages and empowers front line operations leaders and experts, can automate extremely complex and fluid back office processes far more economically than AI based solutions in many, perhaps most Use Cases.

This is not to challenge the wisdom of investing in AI technologies. They are the Future in many, perhaps eventually most business process automation applications, rather they are not ready yet for most demanding processes and even when they become so, may not be cost effective.

Consider the following analogy between three types of anti-tank missile technologies, which are currently used by every major military in the world: Wire, Laser and Radar Guided missiles. None are fully autonomous. They all require a human to acquire the target and make the launch decision. This corresponds to the fact that even various forms of Machine Learning, requires significant human interaction to create, monitor, validate, and fine tune. Still, decision judgment Machine Learning based AI, once sufficient accuracy levels are attained is self-directed, albeit “Fire and Forget”, which can translate into significant capability advantages over systems that require human decision intelligence.

However, Wire or even Laser guided missiles, where humans guide the missiles all the way to the target using sophisticated computer optical or laser technologies makes the guidance easy to provide and almost fool-proof in terms of reliability and accuracy compared either to the more advanced radar guided missiles or obsolete wire guided technology that required significant human skill to acquire and steer the missile all the way to the target, leaving a lot of opportunity for technical failure and human error.

Wow John, that is a great education on Anti-Tank Missiles! What the heck does that have to do with the comparative advantages and disadvantages of various Automation Technologies?

Early generation “Fly by Wire” automation like .Net Macros and first-generation RPA, while a significant efficiency improvement over various non-guided anti-tank weapons (manual processes) are now obsolete. They correspond to Robotic Process Automation or “Humans Augmented by Robots” in the Automation Continuum that follows. Second generation RPA has many enhancements, which improve efficiency and effectiveness. These can be likened to Computer Optical enhance “Fly by Wire” missiles like the ubiquitous TOW; first operational in Vietnam and alive and well in over 45 militaries on over 15 k platforms today. They correspond to Autonomics or “Robots Augmented by Humans”. A variant of the “Fly by Wire” is the laser guided Hellfire. It cost twice as much as the TOW. However, it has significant advantages like twice the range and laser illumination by a platform remote to the launcher, which greatly increases agility and survivability.  This can be compared to extending the ROI range of RPA with OCR and NLP to automate workflows that require unstructured data and Computer Vision that increases automation stability.

There is radar guided version of the Hellfire, which would correspond to Cognitive Computing or “End to End Robots, with Human Oversight”. Again, the advantages of true “Fire and Forget” technologies are manifold; however, they are notoriously hard to perfect in terms of reliability under the full range of operational contingency; something the recent death of a pedestrian by a self-driving Uber car vividly illustrates and are very expensive. The radar guided Hellfire is approximately three times the cost of the laser guided variant.

Sylvan Design Automation Continuum – Notice the resemblance of “Autonomics” or “Robots, augmented by Humans” to the Pareto Automation brand logo that illustrates the article!

Advanced “Fly by Wire” technology has been operational for close to fifty years now and “Fire and Forget” for over twenty. These technologies are all extremely effective tank killers, yet vary greatly in cost per unit. The take home point of the analogy? The older, more simple technologies can continue to have significant usefulness in certain applications long after newer technology arrives on the scene. Because it may not be superior or cost effective in every or even most applications.  Consensus is building it will be the same with Process Automation.

PS – I had a lot of hesitation writing this article. I am not sure it really offers useful information and of course uses another hard to relate to military analogy! Sorry about that really. I wrestled with this for a couple of weeks. The tipping point for me is I just had to have a reason post the illustration. Forget about the text; the picture is worth 10 thousand words!

Source: John Slagboom (Linkedin)-Fly by Wire or Fire and Forget – Augmented Versus Artificial Intelligent Automation?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.