A Close Cousin of Artificual Intelligence, Cognitive System Now Adressing Spend Management, Risk Mitigation and More

There has been much news over past few years on rapid advances in such areas as supply chain analytics and even artificial intelligence (AI). IBM, for example, has released a series of software solutions, including a couple new ones for supply chain, running on its Watson computing system that is built on an AI platform (see Trip Report - NRF 2017.)

A close cousin of artificial intelligence is so called cognitive systems, self-learning systems that use data mining, pattern recognition and natural language processing to mimic the way the human brain works. Most sources place cognitive systems as a subset of AI, and IBM itself describes Watson as a cognitive system.

Supply Chain Digest Says... Jennings says the problems with bad data go beyond just not the best answers. Those faulty answers can create distrust from users even when the answers are right. What do you say? Click here to send us your comments Click here to see reader feedback

Writing in a recent edition of Inside Supply Management magazine from the Institute for Supply Management, Roberta Jennings, a research associate for ISM's CAPS Research, says cognitive systems for procurement are "so sophisticated that they do not just regurgitate or repackage data, but analyze, compare, find correlations and identify trends to the extent that they ‘think.'"

Jennings adds that cognitive systems usually have the following attributes:

• The system can understand data, both structured [basically, in a database] unstructured [e.g., plain text information], and natural language.

• The system can reason, taking into account underlying concepts and relying on various confidence levels.

• The system can learn by accepting feedback and incorporating that feedback into future recommendations.

• The system interacts with users in a natural, intuitive, user-friendly way.

• Cognitive systems go beyond analyzing data at the source and gather data from multiple sources and analyze it together, weighing and considering sources accordingly.

• The system is not merely measuring or checking activity against set parameters. Beyond compliance, it offers solutions that add value to the process.

• With the vast amount of data it accumulates, the system can make correlations that haven't been specifically dictated. In other words, it can estimate what information you're seeking before you know what you're looking for.

• The more information and use the tools get, the more value they can generate. A tool's ability to "learn" and be predictive in its models is a key difference from traditional analytics.

The key factors enabling this revolution of cognitive systems are a wealth of data and emerging Cloud technologies that allow all this data to be captured, shared and accessed by many, often around the globe.

"Combine these with the ability to design a system that can analyze complicated layers of information, and you've got cognitive solutions that enable firms to leap from doing things better to doing better things," Jennings says.

(See More Below)

Jennings cites a number of specific examples of how cognitive systems can improve procurement processes. Here are a few:

Supplier or supply market assessment: For those new to sourcing in a category, a cognitive tool can pull information into an easily digestible format to help identify potential new suppliers, run objective RFI assessments and identify risk factors.

Spend management: Cognitive tools can analyze supplier consolidation opportunities, savings/payment term opportunities and costs of noncompliance.

Supply/materials management: Systems can assess reliability and performance issues on products and services. They can adjust production planning by predicting, for instance, maintenance needs and anticipating bottlenecks relative to inventory/delivery.

Risk management and mitigation: With risk management, sometimes the information that would help you avoid some type of loss is not information that you knew you should be paying attention to. It's unpredictable. Cognitive tools can assess that mass data landscape and identify multitier dependency situations - pinpointing where vulnerabilities could occur.

Contract management: These systems could become a robust repository for contracts and agreements, analyzing contracts for performance versus terms and instigating corrective action, renegotiation or other supplier management activities. Imagine an organization that has acquired another company and all its contracts. Cognitive tools can analyze those fort the acquiring firm's standards and drive contract personnel activity to address areas of concern.

That all sounds great. What are the weak spots of cognitive solutions?

Jennings says the main challenge has to do with data quality.

"It can be extremely challenging to cleanse data, determine weights for various classifications of data and ensure that there is some standardization among measures," Jennings says. "Firms may feel confident about data generated through their own ERP systems, but the situation gets murkier when also feeding external data into a cognitive system."

Jennings says the problems with bad data go beyond just not the best answers. Those faulty answers can create distrust from users even when the answers are right.

But Jennings ends on an optimistic note, writing that "Many experts say that the next 5-10 years will be revolutionary for procurement, thanks to these types of tools," adding that "Ultimately, tools are likely to emerge that satisfy any size organization and budget, so recognizing the role they will be playing in future procurement operations is vital."

What are your thoughts on cognitive systems? Have you seen anything in supply chain or procurement specifically? Let us know your thoughts at the Feedback section below.

Your Comments/Feedback