To understand where analytics interfaces are going, we must first look at a specific failure from the past.
In 1996, Microsoft introduced Clippy, an animated paperclip intended to be an intelligent assistant. Clippy was designed to interrupt. If you typed "Dear," it would pop up and say, "It looks like you're writing a letter. Would you like help?"
Users hated it. They didn't want a presumptuous assistant; they wanted a canvas.
For the last two decades, Exploratory Driven System Usage has defined Human-Computer Interaction (HCI). Users preferred environments where they could pick and choose their path. Designers, in turn, learned to constrain these choices around high-value tasks.
Today, the pendulum is swinging back toward the assistant, but this time, the "paperclip" is invisible, highly intelligent, and deeply integrated.
Embedded within modern analytical tools is a new layer of capability driven by three forces:
Steve Jobs famously said that it isn't the consumer's job to know what they want. Modern interfaces are moving from responding to user input to predicting it. By modeling user behavior, tools can now serve the "right" information before you ask for it. It is the shift from "Here are the options" to "Here is what you likely need." By leveraging machine learning on historical user data, tools can predict needs via pattern recognition.
Modern interfaces draw from a careful—and sometimes manipulative—study of human behavior.