> ## Documentation Index > Fetch the complete documentation index at: https://wiredesk.ai/docs/llms.txt > Use this file to discover all available pages before exploring further. # Analytics > How much your agents handled over 7, 30 or 90 days, how conversations ended, what people asked about, and how they felt. **Analytics** reports on every agent in your workspace over a time range you choose. It answers three questions: how much your agents handled, how it went, and what people asked about. Most figures link to the conversations behind them, so you can go from a number to the transcripts. ## Time range Pick **7d**, **30d** or **90d** in the top right. The default is 7 days. The range covers conversations that *started* in the last 7, 30 or 90 days, counted back from now. The daily chart groups days in UTC. Outcomes, topics and sentiment are written when a conversation ends, usually within about fifteen minutes. Conversations from the last few minutes may not be counted in those charts yet. ## Totals | Tile | What it counts | Click to open | | ------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------- | | **Conversations** | Conversations started in the range, with a per-channel breakdown underneath. | Conversations | | **Leads captured** | Leads saved in the range: people who left their details. | Inbox โ†’ Leads | | **Actions run** | Every tool call your agents made in the range. That includes knowledge searches as well as bookings, API calls, lead and detail captures, messages taken and requests for a person. | Inbox โ†’ Extra fields | | **Human handovers** | Conversations in which a person was asked for, or that ended **Handed to you**. | Conversations filtered to **Handed to you** | | **Answered by a person** | Of the conversations where someone asked for a person, the share where a teammate joined. See below. | Conversations | ### Answered by a person This tile tells you whether anyone actually turned up when a customer asked for a person. * **The percentage** is the number of conversations where a teammate joined after the request, divided by the number of conversations where a person was requested. Joining means replying or clicking **I'll take this one**. * **The line underneath** reads, for example, "4 of 5 ยท typically 3m". That time is the average wait between the request and the first teammate joining. * A teammate who steps into a conversation where nobody asked for a person does not count toward this figure. * A request that timed out and was answered late still counts as answered, with its full wait. If nobody asked for a person in the range, the tile shows **โ€”**. Live phone transfers are not counted here. Those callers are put straight through. ## Conversations per day and channel mix * **Conversations per day** is stacked by channel (Chat, Phone, Email, Browser voice), so you can see where a busy day came from. * **Channel mix** shows the share of conversations on each channel over the whole range. ## How conversations ended Every conversation gets one of four outcomes when it ends. This card shows how many conversations in the range had each outcome and their share of all conversations with an outcome. Click one to see those conversations. | Outcome | Meaning | | ----------------- | ------------------------------------------------------------------- | | **Resolved** | The agent answered what they came for. | | **Handed to you** | A person took over: a live transfer, or a reply from the dashboard. | | **Message taken** | The agent could not help and took a message for you. | | **Abandoned** | They left before the agent could help, or never said anything. | The outcome is decided by reading the whole transcript. A conversation in which a teammate joined is filed **Handed to you** unless it was clearly resolved. A request for a person that nobody answered is not filed as **Handed to you**. A conversation where the customer never said anything is **Abandoned**. More on how outcomes are written: [Conversations](/docs/workspace/conversations#summary-outcome-topic-and-sentiment). ## Where the agent got stuck The five topics that most often ended **Message taken** in the range: the agent could not answer and took a message instead. This is the shortest list of things to teach your agent next. Click a topic to open those conversations. Then use the **Teach** card on the conversation page to add the right answer. See [Fix a wrong answer](/docs/knowledge/fix-a-wrong-answer). ## What people asked about The eight most common topics in the range, with how many conversations each covers. Topics are short phrases written when a conversation ends, such as "appointment booking". Click one to open the conversations list searched for that topic. ## How people felt The customer's sentiment across conversations analysed in the range: **Positive**, **Neutral** or **Negative**. It measures the customer's mood, not the agent's. Analytics covers every agent in your workspace. Your own browser tests of the demo agent are counted too. Only you see those. ## Related * [Conversations](/docs/workspace/conversations) * [Dashboard](/docs/workspace/dashboard) * [Fix a wrong answer](/docs/knowledge/fix-a-wrong-answer) * [Handover](/docs/concepts/handover)