Can you explain the concept of "rule-based" decision-making in AI systems?