I posted on LinkedIn in October 2026 about Google’s new open‑source ambient AI agent, AQuA (Ambient Quality Agent). It runs side‑by‑side with my agents in a Google Cloud project, reads production traces after deployments, and points to the exact file and line range with a suggested fix – but never applies the edit itself.
Each sweep samples up to 1,000 real sessions and grades each one for wrong tool calls, ungrounded answers, and dropped constraints.