3 Reasons the Cognitive Era is Not Yet Upon Us — But it Will Be Soon

Forget digital. The 2020s will be powered by super intelligent, human-like applications that all but replace their creators. This is the dawn of the cognitive era.

At least, that’s what the software market has been saying for the past year or so. Big data, machine learning and artificial intelligence are slowly becoming everyday terms in the business world, and marketers seem hell bent on portraying these technologies as tools in everyday use, too.

But given that most organizations, particularly those in a B2B or supply chain context, have barely come around to adopting even plain old “digital” strategies, cognitive’s penetration in the enterprise is, perhaps, a bit oversold — at least in the procurement and supply chain world.

We’re not trying to ruin the party. Really, it hasn’t started yet. Both solution providers and procurement organizations are still out buying supplies, first. (BTW, could you get ice?)

To understand, here are three examples of how we're still laying the groundwork for the transition from digital to cognitive — and what to expect when it really starts.

IBM Bought The Weather Company

Back in 2015, IBM surprised investors when it announced it had acquired much of The Weather Company, which owned Weather.com and Weather Underground. People made (hilarious) jokes about IBM taking its transition to cloud computing too literally.

But what IBM was really doing was creating a foundation, through the consumption of massive amounts of data about weather, for tools that can help companies adjust their supply chains based on weather-related changes in demand.

The immediate applications are intuitive. Using IBM’s weather data, a logistics category manager could see that a storm was moving in the direction of a pending truck shipment, and either alert the trucking operator to change its route or make appropriate mitigations for a potential shipment delay.

What IBM really wants to do, however, is feed all of this weather data to Watson to improve its predictive capabilities. If all goes according to plan, IBM will be able to produce a forecast that puts your local news station’s five-day forecast to shame.

The same principle applies to procurement. Progressive organizations are feeding their purchase requests and POs through solutions that mine this data, much like weather patterns, for insights about who is purchasing what, how often purchases are being made and how well purchases comply with internal policies.

But while these insights are helpful to procurement, they hardly represent a fully human-like purchasing aide. To get there, we’ll need not just a treasure trove of weather data but also continuous improvement.

Alexa is Listening

The runaway success of Amazon Echo/Alexa is proving that consumers are open to new ways of interacting with technology. But the real foundation for the product's success is that it gets smarter all the time — really, it gets smarter every hour.

The beauty of machine learning is that it can sift through the growing pile of data that users are creating to automatically find new patterns. But before we enter a world where we can converse with a Star Trek-like computer — the original inspiration for the Amazon Echo — software providers need to gather reams of unstructured data on which to run and refine their algorithms.

If Amazon can do this in the enormous field of consumer electronics, processing voice requests for everything from music choices to trivia, you can bet the benefits of machine learning approaches will work on procurement processes, as well.

As just one example, the technology for automatically classifying spend data has existed for a while now but has only in recent years matured to the point where it is actually useful.

Whereas yesterday’s “automated classification engines” could barely achieve 80% accuracy, best-in-class providers have created algorithms that rival human classification at 98% accuracy.

The cognitive era won’t begin in earnest until these mature machine learning-based approaches to spend analysis have had time to ingest and practice creating predictive and prescriptive analyses from spend data.

The Machines Still Need to Learn from Us

Even as IBM hoards weather data and Alexa uses daily interactions to continuously improve itself, it’s easy to overstate what both companies, and procurement solution providers, are actually creating from these initiatives.

The applications of machine learning and artificial intelligence you see today actually emulate human intelligence, rather than exhibit true intellect in the way humans do. In fact, while machine learning-based software can make better predictions and find connections better than we can in many cases, it can also emulate “grave stupidity,” as our own Michael Lamoureux explained over on Sourcing Innovation.

“The reality is that even though some knowledge workers are being displaced, there’s a need for knowledge workers to create, maintain and improve these algorithms … and find new areas in which they can push capability forward,” he wrote.

Procurement professionals will be essential in this effort, helping solution providers spot faulty correlations and pushing capability forward in a way that ultimately increases the importance of the procurement function.

As Lamoureux puts it, “The future will be man and machine, in a delicate dance, and the focus will be on cognitive activities, but mainly on the human side … finding ways to properly apply, and verify, new technology. Weeding out the false positives with intelligence, identifying the false negatives with insight and finding new applications the machines themselves will not.”

Share on Procurious

Discuss this:

Your email address will not be published. Required fields are marked *

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