The Seductive Trap Of Predictive Analytics And What It Won't Tell You

Source: Found here

Increasingly, business leaders are relying on predictive analytics models and AI to plan for the financial futures of their organizations. These models, built from consumer surveys and extensive operational data, provide insights that read as almost clairvoyant, particularly in the shorter term. And while it's tempting to believe AI will help us see around corners, sift through massive data sets, and give us tomorrow's news today, it's important to remember this is only a perceived sense of control.

Denise Worrell , an academically trained futurist and human-centered design expert who leads the Innovation + Transformation practice at Langrand , a creative think tank that helps business leaders solve complex problems, points out that relying on AI and data to show us "truth" can be dangerously misleading. Just like a charismatic but unreliable friend, AI can sometimes give us exactly the answers we want, even if those answers aren't entirely accurate. It's a reminder that predictive models can often provide a limited and narrow vision of the future—one that assumes tomorrow will be an exaggerated version of today, with more growth, more decline, or a continuation of the same trends.

This miscalculation led to major markdowns and profit warnings, which in turn caused Target's stock to plunge by 25% in a single day. Target's struggle to align with considerable shifts in consumer preferences underscores how risky it is to rely too heavily on data models that don't consider the unpredictable nature of human behavior.

"There's a certain comfort in data," adds Worrell. "Numbers feel objective, and predictive models bring a sense of order and certainty. But the reality is that people aren't just numbers. Humans are complex, emotional beings and our behavior doesn't follow a predictable, logical path."

One such example Worrell highlights is the use of AI-based customer service technology. According to a recent Prosper Insights ⁘ Analytics survey, Boomers are the generation least likely to prefer AI-based customer service, especially in personally sensitive applications. For example, when it comes to healthcare matters, Prosper's data shows that Boomers are 3 times more likely to say they prefer talking to a live person over AI, more than all U.S. respondents combined.