Computers are smart, and they’re getting smarter every day. It’s amazing how often I hear about which jobs will be overtaken by machines, and to a large extent, it’s true. Many human functions will be replaced by machines. However, humans will always be smarter—if not faster—than machines, because we have one thing they don’t: empathy. We understand the other person. Therefore, when it comes to making data intelligence pervasive throughout your organization, the humans will guide the way.
Sure, computers learn, but people are—and always will be—needed as interpreters between machines and human needs. If you don’t have humans who can make sense of the algorithm results, the fickleness of customer desires, and the needs of knowledge workers who use the systems that help run your business, that expensive infrastructure you invested in is nothing more than a mass of metal, plastic, and wires.
Understanding the data
To be sure, machine learning algorithms are revolutionizing predictive analytics. Machines—if they’re fed enough data—can predict what’s around the corner. They can also make prescriptions for actions on those events most likely to occur.
But therein lies the rub. What’s enough data, and—more importantly—is it the right data? The right data is bias free—or at least as much as possible. It’s also diverse enough to ensure that the predicted outcomes aren’t skewed. Finally, the right data is appropriate. By appropriate, I mean it’s data that is ethical and intended for consumption—not data that’s gathered by inappropriate means or without consent. You’d think that wouldn’t be an issue, but machines don’t care what data they wrangle.
Humans play a big part in ensuring that machines get the right data. It takes a human to recognize and correct for biases in data. It takes a human to recognize whether or not data sets are diverse enough to produce sufficiently-accurate predictive outcomes. And, finally, it takes a human to ensure that data collection methods are both ethical and informed. That’ll never change.
Understanding the customer
One of the best uses of predictive analytics power is to understand and meet customer expectations. People are fickle, and machines can tease out patterns and trends from seemingly-incomprehensible behavior. That’s a good thing, and it can help you meet customer demands for better intimacy and products that meet needs more effectively. Analytics can also help you predict products customers will want before they know they want them—think iPhones and TiVo.
However, for all their predictive power, machines don’t have hunches. They don’t really know humans with all their irrationality and stubbornness. So, sometimes those predictions and prescriptions go wrong—think Amazon Phone, Facebook Phone, and Google+.
I’m not here to pick on anyone, or any company. My point is that computers, no matter how “smart” they become, can only be as intelligent as the people who interpret the algorithm results and apply models to the real world. Therefore, it’s critical that each model you produce be evaluated in the context of the market experience and collective wisdom of the organization. Your product or service may resonate with your model, but it may flop with actual, living, breathing humans, so you need a human reality check on models you develop.
Understanding the people
Before you get to sophisticated decision-making, however, you have to have intelligent systems that aid you in making those decisions. Those intelligent systems are designed for, and by, humans. The key to these intelligent systems is empathy.
Simply put, systems must be designed with user needs and desired outcomes in mind. Business engagement will drive—and be driven by—intelligent systems design. What I mean by this is if users don’t like a system, don’t understand it, and don’t believe in the results, they won’t use it. If users aren’t consulted and heard—really heard—when systems are designed, you’ll get those unusable systems.
Here’s where the principle of cognitive design comes in. This isn’t actually my bailiwick—it’s my good friend Tyler Rebman’s specialty—but Tyler’s work is excellent, so I’m going to steal it and summarize it. Cognitive design is the process of solving real problems for real people. Its foundation is empathy for people—how they think, feel, and act.
The premise behind cognitive design is that empathy will enable you to really understand users’ needs and discover new opportunities to solve business problems along the way. Cognitive design creates happy users who engage with systems repeatedly and use them to make decisions that are based on better insight to drive improved outcomes and ROI.
Those systems that are created with users in mind will be your most used, and most successful, systems. They will be the systems you use to make data intelligence pervasive in your organization.
It’s the people, always
Fundamentally, until computers become sentient (and I don’t really ever want to see that scenario) humans play a vital role in harnessing computing power to meet human needs. Without people, computers are simply agglomerations of electronic parts. So, it’s critically important, when you’re undertaking the process of embedding intelligence into your organization—of making it pervasive—that humans are at the forefront of the effort. The technology is great, but only insofar as the human element that leverages it.