Support does stand out as one of the strongest use cases for AI right now. That thread's data backs it up—support is running serious per-person Claude spend, right behind engineering, because every ticket means pulling customer history, logs, docs, and past cases.
You're spot on about the shape of the work. It's high-context information retrieval, just like engineering, and AI is great at synthesizing it fast so agents can self-resolve more and reduce low-value escalations.
The direct customer impact you mentioned raises the stakes, which is why the highest-leverage setups treat AI as a reliable co-pilot that surfaces answers while humans own the judgment and relationship.
What AI tools or workflows have moved the needle most for your team?