What do we mean by a cognitive computing application?
The past few months have brought an unprecedented wave of news reports about the impact cognitive computing and AI will have on everything from the health of our economy, the safety of our country, the security of our networks, the way we buy and sell things, the condition of our own health—the list could go on and on.
We sense a shift from as recently as a year ago in the way business executives are viewing the emergence of the cognitive era, or the next-generation intelligent business platform. The conversation on cognitive/AI has gone from a tracking item in the bin of futuristic visions kept on the back burner of the chief innovation officer to a top-of-mind challenge for CEOs and their boards of directors.
Now McKinsey, Gartner and other highly placed influencers have joined the chorus with Ginni Rometty at IBM in declaring that the future lies in digitizing the business, and the core breakthroughs gained from going digital will come from integrating cognitive capabilities.
A bit like the fable
But what is a cognitive computing application exactly? Would you know one if you saw one? And would you have a reasonably intelligent way to differentiate a cognitive application from the applications we are familiar with in this early cloud/big data period? We are now hearing claims from many traditional application niches—like business intelligence, search, marketing, customer support, sales force automation and more—that would have us believe that all those legacy software products are now “AI-driven.” Really? If I add a machine learning component to automate a task within my application suite, have I delivered a cognitive application?
What we read in the press today underscores the diversity and the lack of uniformity in the broad claims for breakthrough business value from the cognitive trend. A bit like the fable of the blind men and the elephant, each new report claims a freshly discovered perspective: that the AI elephant is like a rope, or a waving tree branch, or a pillar, or a solid pipe, or a wall, etc. What we see as a result is a reprise of the fable: The blind partisans can only fall to squabbling about their own versions of the truth. It seems that the conversation is in dire need of a “sighted” observer to make sense of the whole and make peace among the interested parties.
Engaging in the trend
As we have mentioned before in this space, the Cognitive Computing Consortium is making a priority of clarifying what defines a cognitive application and what to consider when it comes time to develop one.
We plan to make our frameworks and guidelines available to businesses interested in engaging in the emergent trend of cognitive computing/AI. We would like industry participants to have resources available to answer top of mind questions. How is a person to calibrate their expectations of what value an AI app can provide? What does one look and feel like? What are the successful options around how to design one? How do I understand what it takes to actually develop one? What kinds of people will need to be involved in the development process if the app is to be a success?
There is solid precursor work in that area, not the least of which was developed in the first round or two of attempts to commercialize artificial intelligence. We have seen the development of knowledge engineering. We can draw on insights from protocol analysis. Important insights can still be drawn from the Carnegie Mellon work on managing the software process. And we are witnessing rapidly evolving innovations rewiring the area of human-computer interaction.
The Cognitive Computing Consortium and Babson College’s Division of Technology, Operations and Information Management have launched a joint research program focused on establishing a standard reference framework for describing cognitive computing applications. The program is led by Susan Feldman of the Cognitive Computing Consortium and Professor Ganesan Shankar of Babson College.
The research is based on an ongoing series of structured interviews with organizations that have created and deployed applications of cognitive computing technologies to one or more use cases within their enterprise.
Fourfold purpose
The purpose of the research is fourfold. First, the program will provide a neutral, systematic baseline of use case data for understanding what is actually being accomplished today through emerging cognitive computing/AI projects.
Second, the research seeks to move business understanding of the cognitive computing trend beyond the level of hype currently rampant across the media, the technology vendors’ marketing messages and the voices of the interested service provider communities by providing a standard reference framework for rigorously characterizing cognitive computing/AI projects. The foundation of that framework is based on the documented ability of those systems to capture business value through innovative leveraging of data, technologies and behaviors, as well as high-level integrations across systems of hardware, software, networks and sensors. We categorize cognitive applications based on each of those parameters.
Third, the research results will be made available to organizations that are seeking solid ground to make decisions about working with the new technology approaches: What reasonable expectations are appropriate for various application classes? What skills, competencies, resources and time frames may be required for such projects? The research program also will discover negative examples from which to characterize failures as well as successes.
Fourth, the research data will include the organizational and technical skills required to develop and foster adoption of cognitive computing applications. Access to the body of knowledge on organizational engagement should help the managers of both current and future cognitive systems projects to understand the scope of a cognitive systems effort in terms of human resources and to arrange the coordination and collaboration of business functions appropriate to the context in which the system will operate.
The interview phase of the research is underway. We invite any organization or individual who has experience with bringing a cognitive computing/AI application to life to share their experience with the Babson research team. Interview identities, content and perspectives remain anonymous. To find out more details on how to participate, email research@cognitivecomputing consortium.com.