Home » Artificial Intelligence » Cognitive Computing and Complexity

Cognitive Computing and Complexity

May 10, 2021

supplu-chain

Although the majority of us wouldn’t like giving up the technologies we rely on to make life easier, there is always a nagging feeling that a simpler life would make us happier. Some people think the simple life is like the one described by the late Henry David Thoreau, who, in his famous book Walden, wrote, “It is desirable that a man live in all respects so simply and preparedly that if an enemy take the town … he can walk out the gate empty-handed and without anxiety.” I tend to agree with author Victoria Moran, who, in her book Lit From Within, wrote, “A simple life is not seeing how little we can get by with — that’s poverty — but how efficiently we can put first things first.” The opposite of simplicity is complexity. I suspect our nagging desire for the simple life is more a longing for a way to deal with complexity. Businesses are no different.

 

Boston Consulting Group (BCG) analysts note, “Businesses compete in a world that is growing ever more complex. Disruptive technologies emerge with increasing frequency. Customers’ needs and demands change at breakneck speed. New competitors are always entering the fray.”[1] They add, “To be successful in today’s complex and fast-changing world, companies must be highly agile and flexible — able to identify opportunities and make informed decisions quickly in order to exploit those opportunities. Those companies that are able to hone that agility in the face of increasing external complexity will emerge with a clear competitive advantage.” Many companies are learning that cognitive computing systems can help them deal with complexity at the same time making them more agile and flexible.

 

Complexity and Complicatedness

 

The BCG analysts assert that many corporate attempts to deal with complexity end up leaving them with a state of “complicatedness.” If that doesn’t sound like much of an improvement, the BCG analyst agree. They insist companies will never be agile and flexible enough until they deal with complicatedness. They explain the failure to deal with complicatedness] is typically the result of one, or several, of the following (perceived or real) causes:

 

Inability to Find Root Causes. The first reason companies fail to deal with complicatedness, according to the BCG analysts is that “the problems and the underlying root causes are difficult to identify and, for the most part, unique to each company.” They add, “There is almost never a single silver-bullet solution, and implementing standard best practices does not solve the problems.

 

Problems are Too Big. When “the problems appear too big or too ‘slippery’ to tackle,” companies often try to find complicated workarounds — which become business-as-usual. The BCG analysts observe they often hear refrains like: “We would need an entirely different culture to do that” and “Once the external environment changes, things will work out again.”

 

Inability to Measure Performance. According to the BCG analysts, challenges associated with complicatedness often aren’t addressed because “the problems are hard to measure, and, therefore, it’s difficult to make the business case to tackle them.”

 

Siloed Responsibilities and Data. BCG analysts note that complicatedness often results when “responsibilities for a given problem are divided among various parties and silos, and no one feels compelled to take ownership of the problem.” They add, “This is particularly emphasized if the problem has in the past proved hard to resolve.”

 

The BCG analysts conclude, “When internal complicatedness is not addressed, tangible value is destroyed.”

 

Dealing with Complexity and Complicatedness

 

Although I agree with the BCG analysts that there are no silver-bullet solutions from complicated challenges, cognitive computing solutions can help with most of the challenges noted above. The first challenge — the inability to find root causes — requires the use of analytics. Cognitive computing systems — like the Enterra Cognitive Core™, a system that can Sense, Think, Act, and Learn® — have advanced analytics packaged embedded in them. George Karapalidis (@gkarapalidis), head of data science at Vertical Leap, explains the purpose of the four types of analytics.[2] They are:

 

1. Descriptive analytics. Descriptive analytics can help discover what happened in the past. As Karapalidis puts it, “Before we learn where to go, we need to know where we came from. That’s the key question descriptive analytics solutions tackle.”

 

2. Diagnostic analytics. Diagnostic analytics can help explain why something happened. Karapalidis writes, “Diagnostic analytics tools help you uncover the root cause of some problems.”

 

3. Predictive Analytics. “Predictive analytics,” writes Karapalidis, “‘joins the dots’ between the accumulated and analyzed data points, conveying what and why something happened, into models suggesting what can happen next. It indicates the probability of certain outcomes with high accuracy and takes the guesswork out of your decision-making process.”

 

4. Prescriptive analytics. Prescriptive analytics informs you what you should do to achieve a particular outcome. It’s a type of analytics made possible by the emergence of cognitive computing technology. Karapalidis notes, “Prescriptive analytics is yet to move from the margins to the mainstream. It’s an emerging area of analysis attempting to answer the complex question of ‘what actions to take if I want to get outcome A?’ Prescriptive tools come up with multiple future outcomes based on your current/past actions; match those futures with your goal and advise you on the action you need to apply.”

 

As Karapalidis notes, diagnostic analytics can help discover root causes of problems if (and it’s a big if) you have the right data and the right model.

 

Dealing with problems that are “too big” generally requires better collaboration between multiple stakeholders. Because cognitive computing systems can deal with many more variables than older computing systems, better collaboration can be undertaken. This is true for both internal and external collaboration. Because cognitive computing systems can integrate various types of data sources, they can also help with the challenge of dispersed responsibilities and siloed data. Great collaboration requires stakeholders to have access to the appropriate data — a single source of truth. By helping to break down data silos, cognitive computing systems can assist business leaders in their efforts to create corporate alignment. Cognitive computing systems can also help ensure corporate planning efforts are in alignment by implementing concurrent planning solutions.

 

Although it may sound like I’m selling cognitive computing as a silver-bullet solution, as I noted above, I’m not. Cognitive computing is simply a tool in a larger corporate kit that is needed to address complexity and complicatedness. As the BCG analysts conclude, “Such problems are not intractable, however. The solution lies in gaining a thorough understanding of the system at work. With that knowledge, a company can design a combination of interventions that address the root causes and modify the system in a way that instills and supports the desired behaviors. This is easier said than done, of course.” They suggest companies need tools that address eight dimensions that drive personal and corporate behavior: 1) Strategy and transformation agenda; 2) Structure; 3) Activities and roles; 4) Processes and systems and IT; 5) Decision making; 6) Performance management; 7) People and interactions; and, most importantly, 8) Leadership. Cognitive computing can help with things like improving processes and decision making; however, other tools are required to address the other challenges associated with complexity and complicatedness.

 

Footnotes
[1] Jaap Backx, Christoph Hilberath, Reinhard Messenböck, Yves Morieux, and Henning Streubel, “Mastering Complexity Through Simplification: Four Steps to Creating Competitive Advantage,” bcg.perspectives, 16 February 2017.
[2] George Karapalidis, “Examining the four types of big data analytics,” The Drum, 18 December 2018.

Related Posts:

Full Logo

Thanks!

One of our team members will reach out shortly and we will help make your business brilliant!