Cognitive computing. Machine intelligence. Smart machines. These are just a few of the dozens of phrases that were created as part of rebranding efforts to cut down on the fear and hesitation that artificial intelligence (AI) has inspired in the past. But despite past expressions of reluctance in regards to artificial intelligence, AI is back and it’s here to stay.
Recommendation systems are everywhere. Companies are clamoring for predictive and prescriptive analytics engines. And a day doesn’t go by without news of another learning or advanced reasoning system that is executing on a task better than humans.
The emergence of these intelligent systems and the ever-increasing hype around them has led to companies’ trying to figure out their own AI strategy. The AI category is vast and broad – there are many solutions to solve many problems, but the large and growing number of options creates a new problem: Companies are challenged to understand which solutions truly align with their unique needs. Companies know that they need to adopt intelligent systems or run the risk of being left to stagnate, but a lack of understandable value propositions or well-articulated explanations of the technology can leave executives scratching their heads when faced with deciding on a strategy.
If AI is to be applied properly in the enterprise, this blind adoption strategy will not serve us well. No one should purchase software on the basis of hype and fear. If you do, you are destined to fail before the full potential and return on investment can even be explored.
A New Environment
So, why are technologies that faltered in the past working well today? What happened? It was not a change in the technology but rather a change in the environment. Primarily, it was a change in the data.
The intelligent systems that are succeeding today are built on a foundation of massive data sets that simply did not exist in the past. Google’s Deep Learning, IBM Watson, and the assorted recommendation systems presenting us with suggestions about books and movies that we are likely to be interested in all are driven by huge data sets. It is the data of our work and world. It is the data that are directly related to the tasks that these systems are trying to address. And, even more important, it is the data tied to these processes that deserve a closer look as a potential area for automation.
As companies mature and expand, a suite of best practices is often developed. For some companies, this may include methods and approaches that aren’t just best practices, but proprietary processes that may even define IP and bolster the organization’s brand. These specific approaches, analyses, and methodologies may even become be part of the company’s team onboarding process and define the very reason that your clients hire you.
Of course, if there are well-understood methodologies, they are usually linked to well-defined data sets. Standard practices often rely on stable, non-shifting data. Otherwise, the repeated methodologies are difficult to clone.
Such processes and their data provide perfect opportunities for automation. The better a company understands what it does and how to train its staff to operationalize that understanding, the more likely that it is a process that can be codified, and then that “code” can be used to drive an intelligent machine.
So as you assess where intelligent systems may improve your business operations, ask yourself, “What do we know how to do really well?” and “What data do we use to do it?” Your answers will determine where to start when you begin to build your strategy as well as what parts of your business potentially can be automated at scale. And after you have done so, the next question that you may want to ask yourself is, “How will we use our newfound free time to improve other parts of the business?”
Kris Hammond, Ph.D., is co-founder and chief scientist at Narrative Science and professor of computer science at Northwestern University. Prior to joining the faculty at Northwestern, Kris founded the University of Chicago’s Artificial Intelligence Laboratory. His research has been focused primarily on artificial intelligence, machine-generated content, and context-driven information systems. Kris currently sits on a United Nations policy committee run by the United Nations Institute for Disarmament Research (UNIDIR). He received his Ph.D. from Yale.
Subscribe to Data Informed for the latest information and news on big data and analytics for the enterprise.