> For the complete documentation index, see [llms.txt](https://learn.withessential.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://learn.withessential.com/administration/using-the-essential-assistant.md). # Essential Assistant The Essential Assistant lives in a floating bubble in the corner of every signed-in page of the web app. Click the bubble to open the panel, and click it again (or the X in the top bar) to close it. You can drag the bubble to reposition it. Inside the panel, the Ask Essential tab is where you chat; the Talk to a human tab is where you reach Essential's team for support and setup requests. To start fresh, open the chat list and click New chat. Your past conversations are saved there, sorted by most recent, and you can search them by typing in the Search chats box or reopen one by clicking it. To remove a conversation, delete it from the list and confirm; this can't be undone. On a phone the panel is only as wide as the screen, so the past-chats list is not shown beside the conversation. Tap the menu button in the top-left of the panel to slide the list over the conversation, and it closes again as soon as you pick a chat or start a new one. When you reopen a long conversation, scroll to the top of the chat to load older messages. You can keep scrolling upward to move farther back. The assistant can also refer back to files shared earlier in the same conversation, so you don't need to attach a file again just because it is outside the messages initially shown. The assistant is automatically aware of the page or record you're viewing, so you can ask about "this quote," "this customer," or "this inspection report" without restating which one. The input placeholder reflects that context. It also knows you are asking from the web app, your role, and whether you can sign in to the mobile app. A how-to answer therefore names the app a task is done in: if you ask how to clock in to a visit, it tells you that happens in the mobile app, and if you do not have mobile access it says a technician has to do it. Technicians get the same assistant for how-to questions on the mobile app's Support tab; see [Get help in the mobile app](/field-operations/mobile/get-help-in-the-mobile-app.md). Press Enter to send your message. To write more than one line — a short list, a few separate questions — press Shift + Enter to start a new line. The line breaks you type are kept, both in the message box and in the sent message shown in the conversation. You can also give the assistant a file to look at — a field photo, a PDF report, a spreadsheet of devices. Click the paperclip in the message box to pick files, or drag them straight onto the chat panel, up to ten per message. See [Attaching files to the Assistant](/administration/using-the-essential-assistant/attaching-files-to-the-assistant.md) for the supported formats and size limits. When voice is available, click the microphone and speak. The assistant sends your request after you stop speaking, reads its reply aloud, and then opens the microphone for your next request. Use the arrow to send immediately or **Stop audio** to end the voice session. If you do not start speaking, a warning appears after 10 seconds and listening pauses after 20 seconds. Closing or hiding the assistant also pauses listening. A conversation that has wrapped up may pause automatically after the reply. If spoken replies are unavailable, the microphone still supports dictation: stop recording, review the transcription in the message box, and send when ready. You can give the assistant its own name and personality from the Customize agent option, which shapes how it writes. When voice is available, choose a voice in Customize and select **Play sample** to hear it. Each voice reads the same text so you can compare them. Select **Save** to use your choice for spoken replies, with Jade used when no voice has been saved. Playing a sample does not save your changes. Until you save a customization, it uses the default Quinn name and personality shown there. The assistant streams its replies as it works, and only some of its abilities make changes; many requests are read-only. While the assistant is working, the Send button becomes a Stop button. Choose **Stop response** to cancel the current reply. The stopped turn remains in the conversation as **Response stopped**, and you can send a new request right away. If a reply cannot finish, the failed turn remains in the conversation as **Agent failed to respond. Please try again.** You can send a new request without deleting the failed turn or restarting the chat. When the assistant proposes a change to a draft quote or invoice, it arrives as a proposal card with Approve and Reject — nothing changes until you approve it. See [What the Essential Assistant can and can't do](/administration/using-the-essential-assistant/what-the-essential-assistant-can-and-cant-do.md) for the full list of what it can and cannot change. > **Where to find it:** The floating sparkle bubble in the corner of any signed-in page. Click it to open the assistant panel, which has an Ask Essential tab and a Talk to a human tab. ## Related guides * [Editing a quote or invoice through the Assistant](/administration/using-the-essential-assistant/editing-a-quote-or-invoice-through-the-assistant.md) * [Editing an inspection report through the Assistant](/administration/using-the-essential-assistant/editing-an-inspection-report-through-the-assistant.md) * [What the Essential Assistant can and can't do](/administration/using-the-essential-assistant/what-the-essential-assistant-can-and-cant-do.md) --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://learn.withessential.com/administration/using-the-essential-assistant.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.