Cognitive computing - Part 1
Cognitive computing and the evolution of knowledge work

Potential impact on knowledge work

Anyone who owns a smart phone has experienced the power of machine learning. For example, have you ever looked down at 6 p.m. and seen a message about the estimated drive time and traffic conditions on your route home? That may have left you scratching your head since you never told your phone where you work, where you live or what time you leave every day—and you probably never explicitly programmed your phone to provide a traffic update at a given time. Instead, the software in your phone analyzed GPS data about your routines and movements to predict that you probably will be leaving work at a particular time and probably want to know that there’s a huge accident blocking your normal route.

But the possibilities for machine learning and cognitive capabilities extend far beyond those targeted applications. IBM identifies three broad capability areas for cognitive computing:

• engagement—programs that function as expert assistants by responding to questions and providing targeted information and answers;
• discovery—programs that identify previously undetectable patterns and relationships within large, diverse data sets; and
• decision—programs that make recommendations and decisions by analyzing complex, evolving pools of evidence and assigning confidence scores to various options.

Obviously, all three areas have the potential to augment and transform knowledge work. Some applications—such as systems that analyze vast medical data sets to identify previously undiscovered statistical relationships among symptoms, diseases and outcomes—will aid and supplement human decision making. Others—such as systems that look at an individual’s financial situation and goals to provide investment planning advice—may take the place of some human experts.

The clearest applications for KM are intelligent systems that can process multidimensional information about an employee such as what that person is working on, his or her level of expertise, whom they interact with most often, and other contextual factors in order to recommend possible answers, solutions or actions to pursue—and continue to take in data and generate more refined results over time. Even if a problem still requires the tacit knowledge and reasoning power of a human (and many will), machine learning may be able to surface the most applicable content, data sets or experts to support decision making in a given scenario. In either case, the hope is that employees will be able to spend less time searching for knowledge and more time learning from and using it.

The intersection of cognitive computing and KM

After discussing many potential scenarios, APQC’s KM Advanced Working Group identified six areas of KM where it believes cognitive computing has the most potential:

• search and discovery—The possibilities for cognitive computing to enhance enterprise search and discovery processes are enormous. Current search applications can traverse unstructured data and perform some automatic tagging, but the capabilities are often limited. As search functionality evolves to incorporate cognitive capabilities, it will be able to provide more comprehensive answers, even when that involves interpreting images or videos or combining information hidden in multiple documents.
• content curation—Machine learning can automate content curation by using algorithms to look at different types of data, find similarities between sources and cluster them in logical groups. The content then can be tagged accordingly and displayed both proactively and on demand in response to search queries. The goal is to improve the efficiency (and cost) of curation processes while eliminating the “misses” where relevant content exists, but the system fails to connect users to the best resources.
• expertise location—Cognitive systems have the potential to scan numerous disparate sources of information in order to identify subject matter experts or people who work on specific topics. Such systems can improve the quality of colleague recommendations, reduce the time required to find help and answers, and uncover hidden pockets of expertise—all while relieving individual users from the burden of creating profiles listing their skills, experience and interests.
• data-driven visualization—Typically, people visualize for three reasons: to explain, explore or exhibit data. Cognitive computing facilitates those functions with automation, increased processing power and incredible efficiencies in speed. In the age of big data, visualization can be used to display content and visually map data and knowledge to allow additional connections between topics and insights to be discovere
• lessons learned analytics—Although many organizations are successful at capturing lessons from projects and programs, they often struggle to truly learn from past experience and apply lessons in the future. Programmed correctly, a cognitive system could analyze existing lessons learned databases as well as other project documentation not specifically called “lessons learned” for patterns or trends that could point to opportunities for improvement. Relevant insights then could be delivered to individuals and teams based on the stage and type of project they are working on.
• the digital sidekick—Most people are already familiar with current-generation intelligent personal assistants such as Apple’s Siri, Google Now and Microsoft’s Cortana from their smart phones and other devices. However, more advanced versions have the potential to transform how employees interact with enterprise knowledge. If turbo-charged by the capabilities of cognitive computing, a digital program could serve as a mentor, guide or sidekick to an employee both in the office and on the go.

Experts have been hinting at cognitive computing’s promise for several years, but few knowledge management programs have really grappled with the details of how cognitive technology will affect the flow of knowledge inside their organizations. Future installments in this series will delve into cognitive computing’s impending impact in the six areas, as well as potential challenges and advice for knowledge managers looking to harness these technologies.