> ## Documentation Index > Fetch the complete documentation index at: https://docs.rippit.com/llms.txt > Use this file to discover all available pages before exploring further. # Monitor your AI agent > Find where your AI agent gives wrong answers, fails to resolve, or hands off because it had no answer. Rippit reads every conversation your AI agent has, so it can tell you where the bot is failing and what it needed to know. Resolution rate and CSAT can't tell you that: a customer who gives up counts as resolved, and most customers never leave a rating. ## Ask Rippit ```text wrap theme={null} Look at the last two weeks of conversations our AI agent replied to. Where did it give a wrong answer, fail to resolve the issue, or hand off because it had no answer? Group what you find by topic and customer segment, and for each handoff, tell me what our human agent said that the bot did not. ``` To answer, Rippit can: * Tell the bot's replies apart from your human agents' replies in the same conversation. * Read each thread in full, including what your human agent said after the handoff, which is often the answer the bot should have given. * Judge every bot reply on its own: correct, wrong, no answer, or a customer who gave up. * Break the findings down by your conversation attributes, such as tags, CSAT, channel, customer segment, or product area. Ask Rippit to support each finding with conversation links, so whoever owns the bot can check the evidence before changing anything. To get the list every week without asking, turn the question into an [Agent App](/plg/agent-apps) that posts it to Slack or email. Rippit tells you what to fix; it does not change your AI agent's knowledge or your help center.