> ## 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. # Ask Rippit questions > Use chat to analyze your customer conversations, build reports you keep, and learn the product. Chat is where most work in Rippit starts. Ask a question in plain language and Rippit picks the data, runs the analysis, and answers with the work attached, so a chart or a worksheet comes back alongside the words. Click **New chat** in the left sidebar to start one. Past chats are listed under **Chats**. ## How to talk to Rippit Talk to it like a colleague who has read every one of your customer conversations. Full sentences, not keywords. The one habit worth forming: ask for the judgment you want, not the query you think Rippit needs. "Why are customers frustrated about billing?" is the question worth asking. "Count the tickets tagged billing" is just a lookup, and the tag was probably applied inconsistently anyway. A few things that help: * **Say what slice you mean.** "Last 30 days" or "enterprise accounts only" beats leaving it open. * **Stay in the same chat.** Rippit keeps the context, so follow-ups can be short: "now split that by channel." * **Name the table you mean.** If you have more than one, say which, rather than letting Rippit guess. * **Drop in files.** Images, PDFs, text files, and CSVs can go straight into the conversation. ## What comes back An answer is prose plus whatever Rippit built to get there. Charts, worksheets, canvases, and Agent App drafts appear in the **Artifacts** panel beside the conversation and stay in your workspace afterward. **Show sources** reveals what the answer was based on. If you type while Rippit is still working, your message queues rather than interrupting. **Stop** ends a response that is heading the wrong way. Chats name themselves from your first message, so you can find them later without naming anything. ## What to use it for ### Find out why customers are contacting you The questions Rippit is built for are the ones that need judgment about what was actually said, not a filter over ticket metadata. * "Why are customers frustrated this month?" * "What are the top five reasons people contact us, and which is growing fastest?" * "Which issues did we fail to resolve on the first reply?" Rippit reads whole conversations to answer these, so you get what happened in the thread, not how someone tagged it. ### See what is costing you revenue The same reading works on anything your customers said, not just support tickets. If your workspace brings in sales call transcripts, Rippit analyzes those the same way. * "What objections come up most in the calls that did not close?" * "Which competitors get mentioned, and how do our reps handle it?" * "What reasons do customers actually give for cancelling, and how many are about price?" With a CRM connected, such as Salesforce or HubSpot, it can carry that back to accounts: "Which accounts renewing this quarter have complaints we never resolved?" ### Decide what to fix or build next Your conversations are a running record of what the product gets wrong. Rippit turns it into a ranked list instead of anecdotes. * "What are customers asking for that we do not offer?" * "Which bugs drive the most contacts?" * "Did complaints about the new checkout flow drop after the fix shipped?" ### Build something you keep Ask for a breakdown, a report, or a table and Rippit builds it as a real artifact in your workspace rather than a wall of text in a chat. Ask it to add a column classifying every conversation by theme, urgency, or root cause, and that column becomes something you can sort, filter, and chart later. ### Pull in context from outside Rippit If your workspace has connected MCP servers, Rippit can reach them in the same conversation where it is analyzing your tickets, so it can line up a spike in contacts with what your team was saying in Slack that week. See [Rippit Agentic Chat + MCP Client](/agentic-chat/overview). It can also search the web for facts that live outside your workspace, like vendor details or industry benchmarks, and cites what it used. ### Learn how to use Rippit Rippit answers questions about itself, so chat is usually faster than going to find documentation. * "What can you actually do?" * "What is the difference between a table and a worksheet?" * "How do I classify every conversation by root cause?" * "Why does my table only have 30 days of conversations in it?" When the answer depends on your workspace, Rippit reads your actual setup rather than describing the product in general, so that last question gets an answer about your table. Anyone in your workspace can use chat. If you do not see **New chat**, contact Rippit support.