Primer: Make sense of cognitive computing
- 05 June, 2017 20:00
If you’ve been seeing the word “cognitive” a lot lately, you’re not alone. And if you’re confused about exactly what it means from an IT and business perspective, you’re not alone in that either.
To help provide some clarity around the cognitive concept and what it might mean for your organization, I’ve put together this primer.
What does ‘cognitive’ mean in the context of computing?
Cognitive computing uses technology and algorithms to automatically extract concepts and relationships from data, understand their meaning, and learn independently from data patterns and prior experience—extending what people or machines could do on their own, says Paul Roma, chief analytics officer at consulting firm Deloitte Consulting.
There are three main ways cognitive computing can be applied today, Roma says:
- Robotic and cognitive automation to automate repeatable tasks to improve efficiency, quality, and accuracy.
- Cognitive insights to uncover hidden patterns and relationships to identify new opportunities for innovation.
- Cognitive engagement to drive customer actions by delivering hyperpersonalization at scale.
How is cognitive computing different from AI?
Deloitte refers to cognitive computing as “more encompassing than the traditional, narrow view of AI [artificial intelligence],” Roma says. AI has been primarily used to describe technologies capable of performing tasks normally requiring human intelligence, he says.
“We see cognitive computing as being defined by machine intelligence, which is a collection of algorithmic capabilities that can augment employee performance, automate increasingly complex workloads, and develop cognitive agents that simulate both human thinking and engagement,” Roma says.
Vendors use different names to describe these technologies, says Dave Schubmehl, research director, cognitive/AI systems and content analytics at research firm International Data Corp. (IDC). “Some folks use the name of the types of algorithms to describe the platforms,” he says, such neural networks, also known as deep learning, or machine learning.
“These are some of the key ingredients to building these intelligent applications,” Schubmehl says. “Some use the generic term in the field for this type of application: artificial intelligence. Yet another group uses the phrase coined by IBM researchers when they were working on Watson for the ‘Jeopardy’ challenge: cognitive computing. In all of these cases, the terminology is more or less describing the same field of effort.”
How widespread will the use cognitive computing and AI be over the next decade?
The technology will be “extremely common as an aspect of applications,” says Whit Andrews, vice president at research firm Gartner. The firm has predicted that by 2018 30 percent of interactions with technology will be through “conversations” with AI. And by 2020, AI will be a top-five investment priority for more than 30 percent of worldwide CIOs, Gartner estimates.
With the confluence of exponential data growth, faster distributed systems, and smarter algorithms, cognitive computing “is on a path towards increased permeation across business processes in the areas of robotic and cognitive automation, cognitive engagement, and cognitive insights,” Deloitte’s Roma says.
What are examples of cognitive technology in the enterprise today?
Although much of the promise of cognitive technology might lie in the future, some organizations are deploying cognitive tools already.
Companies are using cognitive systems for product recommendations, pricing optimization, and fraud detection, Schubmehl says. Organizations are also using conversational AI platforms (in the form of chatbots) for automated customer support, automated sales assistance, and decision augmentation, he says.
In health care, Roma says a leading hospital that runs one of the largest medical research programs in the United States is “training” its machine intelligence systems to analyze the 10 billion phenotypic and genetic images stored in the organization’s database.
And large health benefits company is pursuing a cognitive strategy that will encompass automation, engagement, and insights to ultimately streamline and enhance engagement with customers, Roma says. “They are focused on applying cognitive insights to the claims process to provide claims reviewers with greater insight into each case for a more comprehensive assessment,” he says.
In financial services, a cognitive sales agent uses machine intelligence to initiate contact with a promising sales lead and then qualify, follow up with, and sustain the lead. “This cognitive assistant can parse natural language to understand customers’ conversational questions, handling up to 27,000 conversations simultaneously and in dozens of languages,” Roma says.
The most common uses are for performing advanced classification— such as routing people and needs to the best workers to fulfill requirements—and for predictive analysis, such as knowing the best way to promote a product to a buyer, Gartner’s Andrews says.
What are some of the ways cognitive might work in an enterprise?
Organizations will use cognitive/AI technologies to automate business processes, streamline contract analysis and renewal, communicate, sell and support customers, and even automate delivery and resupply of stock in their businesses, IDC’s Schubmehl says.
One application of this added intelligence will be to enable more precise decision-making for business functions such as sales and marketing. “We expect organizations to make their decisions highly specific,” Gartner’s Andrews says. “It’s easy to develop promotions for all customers today; in the future we expect to see true personalization. We [also think] it will allow for more effective autonomous vehicles and transport systems.”
The possibilities for cognitive are boundless, says Bret Greenstein, IBM’s vice president of Watson Internet of Things Platform. “Cognitive capabilities will expand in their understanding of all different types of information—sights, sounds, emotions, etc.—and will develop more sophisticated ways of learning from us and from data to better support every job,” he says. “The idea in the future would be that all jobs are enhanced with cognition.”
Which industries are most likely to be affected by the emergence of cognitive technologies?
The financial services sector today is showing the greatest interest in cognitive technologies, Andrews says. “We see elevated levels of inquiry, searches on our website, and social media signals from and about financial services and AI,” he says. “Data in financial services is of greater volume and quality than it is in most verticals. That makes it ripe for advanced analytical strategies.”
But cognitive computing’s potential has applicability across just about every major industry that relies on data-driven decision-making to improve outcomes; where efficiency and accuracy gains can be realized through the automation of some processes; and where mass-consumer personalization at scale is required, Deloitte’s Roma says.
“Any industry where data is gathered and can be used to gain insights will be affected,” IBM’s Greenstein adds. “Cognitive technologies can open new markets, deliver efficiencies, and deliver competitive advantage by delivering real-time insights that are actionable.”
In sectors such as financial services, health care, manufacturing, legal, and public sector, competitiveness is increasing their dependence on “finding that needle in the haystack faster so they can improve their quality and timeliness of their actions,” says Brian Cowe, a senior product manager at Hewlett Packard Enterprise.
What are some of the major challenges with cognitive computing?
Some of the biggest challenges revolve around transparency of decision-making based on data, as well as its trustworthiness, IDC’s Schubmehl says. “Organizations also have to be careful of providing too much information and/or decision-making that the product or service becomes unattractive to the consumer or user,” he says.
To gain the biggest benefit possible from cognitive technologies, enterprises need the ability to connect and combine all their internal data with that of public data, Greenstein says.
“This presents a challenge, given the volume of data created each day in any given industry and the fact that it is often siloed in various locations,” Greenstein says. “Add to that the fact that up to 80 percent of business data is not searchable. This is why it is so critical that enterprises go through a digital transformation, embracing the data of their own business and the world around them.”