The Bosses Using AI to Guess Who Quits Next — predictive…

Your manager might know you're about to quit before you do. That's not a paranoid fantasy—it's a product you can buy right now. A fast-growing category of workplace software is quietly turning the art of "reading the room" into a math problem. These tools, usually labeled "predictive intelligence" or "people analytics," promise to tell employers which employees are about to walk out the door, who's about to underperform, and which teams are one bad week away from falling apart. And American companies are buying in. **What is it, in plain English?** Think of it like a credit score, but for your job. Instead of a FICO number, you get a "flight risk" score. The software pulls together everything your employer already knows about you—how often you log in, how many meetings you skip, how long since your last raise, how your last performance review went, whether you've been quietly browsing LinkedIn on company Wi-Fi—and runs it through a model that spits out a probability. Some tools even flag which managers are most likely to lose their teams. The pitch to executives is simple: turnover is expensive. Replacing a mid-level employee can cost six to nine months of their salary once you add recruiting, onboarding, and lost productivity. If software can spot the warning signs two months early, a well-timed conversation or a raise might keep them. That sounds reasonable. **So why does it feel like a spy movie?** Because the same data can be used two ways. The optimistic version is an early-warning system for retention—a manager notices your engagement is dropping and asks if you're okay. The cynical version is a hit list. If the model says you're 80% likely to leave, why promote you? Why put you on the big project? Why give you the raise? You become a bad investment before you've done anything wrong. There's also the accuracy problem. People are not weather systems. A model trained on last year's quitters might totally misread this year's workforce, especially after layoffs, return-to-office mandates, or a hot job market. And once you start managing people based on a score, you risk creating the very disengagement you were trying to catch. Nobody likes feeling watched. **What this means for your wallet and your week** You probably can't opt out. Most of these systems run on data your employer already collects, and in most U.S. states, that's legal. But you can protect yourself. Keep your resume current even when you're happy. Document your wins somewhere you own—not just the company drive. And if a boss suddenly gets very interested in your five-year plan, don't assume it's pure kindness. It might be a model flagging you. **The bottom line** Predictive intelligence isn't inherently evil. Used honestly, it could make workplaces fairer by spotting burnout before it becomes a resignation. Used lazily, it turns your career into a number you didn't agree to be graded on. The difference isn't the algorithm—it's whether your boss treats the score as a question or a verdict. **Our take:** You can't stop your company from running the numbers on you, but you can refuse to live inside their spreadsheet. Keep your options warm, your skills sharp, and your documents in your own folder. The best predictor of your next move should always be you.