Cognitive computing - Part 1
Cognitive computing and the evolution of knowledge work
Cognitive computing will be as disruptive to organizations in the next decade as social media was in the last—and perhaps even more so. In fact, Gartner has predicted that the smart machine era will be the most disruptive in the history of IT.We are on the brink of a paradigm shift involving the fundamental human processes that guide information discovery, insight extraction, problem solving and decision-making.
Cognitive computing, machine learning and predictive analytics will permeate every aspect of our lives and radically transform how we learn and interact in our digital lives. Content management, collaboration and the entire search experience will evolve to become more automated, seamless and personalized. The result is that we will rely on computers even more heavily than we do now while forging increasingly complex—even intimate—relationships with them.
And the future is already here, at least for a handful of early adopters. Cognitive capabilities are starting to play a significant role in industries as diverse as healthcare, software, financial services and oil and gas exploration. In a 2016 poll of APQC’s knowledge management audience, 9 percent of participants said their organizations are already using cognitive technology, with another 17 percent planning or seriously considering cognitive applications (see chart on page 21, KMWorld January 2017, Vol 26, Issue 1 or download chart). This is the moment for organizations to start honing their strategies and preparing for the coming changes in how employees create, share, access and apply institutional knowledge.
The implications for knowledge management have not yet been fully articulated, but KM professionals need to be informed about what cognitive computing can do and how they will need to adapt. To explore the impact of cognitive technology on KM capabilities and programs, APQC convened its 9th KM Advanced Working Group, which brought together experienced KM practitioners from APQC, Deloitte, EY, NASA, Pfizer, and the U.S. Army Training and Doctrine Command. Together, this group investigated the opportunities and potential use cases for leveraging cognitive computing to manage and deliver enterprise knowledge.
What is cognitive computing?
Cognitive computing is an umbrella term for computerized models designed to mimic human thought processes. Cognitive systems can, according to Deloitte, “process information far beyond human capabilities, identifying patterns and providing potential solutions that humans might never recognize through traditional analysis.”
A true cognitive system combines machine-learning algorithms (which allow software to improve through experience) with advanced data mining capabilities (which facilitate the identification of patterns and relationships within large data sets) and natural language processing (which enables the computer to derive meaning from human speech and respond using language like a human would). Perhaps the best-known cognitive solution is Watson, IBM’s technology platform that uses natural language processing and machine learning to reveal insights from large amounts of unstructured data.
The result is a computer capable of sifting through vast quantities of structured and unstructured content to provide personalized, intuitive responses to a human user’s search or question. And as cognitive models expand to incorporate new data and information, they become progressively more comprehensive and accurate.
Basis for the emerging cognitive revolution
Traditionally, computers have excelled at queries with clearly defined right and wrong answers. Software has become much more sophisticated over the past two decades, and today’s cognitive systems and machine learning applications approach questions more dynamically. That allows them to tackle more complex problems, refine their responses based on influxes of new information and detect the best options among numerous viable possibilities. Those capabilities may eventually allow cognitive systems to respond to the type of multifaceted, context-driven inquiries typically reserved for human intelligence.
APQC’s KM Advanced Working Group hypothesized that three phenomena will propel the coming surge in cognitive adoption.
The first is the sheer amount of data available to both organizations and employees. According to Singapore-and India-based Aureus Analytics, 90 percent of the data in the world was created in the last two years, with the total volume expected to grow 40 percent each year as the Internet of Things comes online. Migration to the cloud provides ready access to a lot of that data, but it is impossible for the human mind, even when augmented by traditional computer programs, to process or make sense of so much information. Cognitive systems are needed in order for organizations to get their arms around big data and detect patterns and insights among the chaos.
The second factor relates to the technology itself. Recent advances in raw computing power and analytical models are making it possible for systems to traverse vast data sets (including unstructured and rich-media content) and return meaningful results in near real time. Semantic analysis opens up qualitative data for analysis, and machine-learning technology enables systems to interpret a user’s context and pose follow-up questions to further refine their responses. And finally, developments in natural language processing and user interfaces have improved human-computer interaction and paved the way for closer, more natural partnerships between man and machine.
The final factor is the high cost of knowledge work. Many organizations see their workforces as their single biggest investment—and one that isn’t being leveraged as efficiently as it could be. According to the IDC, knowledge workers on average spend almost 20 percent of their time looking for internal information. A more intuitive search capability that encompasses internal, external, structured and unstructured content and provides highly targeted answers (instead of lists of pages or documents the user must sift through) has the potential to save organizations billions of dollars while freeing up employee time for more value-added activities.