AI Recommendations Aren't a Click Curve You Can Extrapolate

For two decades, search visibility was something you could model: know your rank, forecast your traffic. AI-generated recommendations break that bet entirely. These systems are non-deterministic by design — the same question produces different answers, different brands, and different citations from one run to the next, with studies showing less than a 1% chance of getting the same brand list twice. Yet AI search is the fastest-growing and highest-converting discovery channel on the web. This deep dive explains why AI recommendations resist forecasting, what the 2026 data actually shows, and how to shift from gaming a rank to managing probabilities.