--- title: How to ask the Agent about your tasks, chat and documents url: https://tideswell.xyz/docs/analyst/ask-about-your-work description: Ask the Agent what is late on a project, what a channel agreed or what a document says, answered from what you can see. --- > Documentation index: fetch https://tideswell.xyz/llms.txt to discover every page before exploring further. # How to ask the Agent about your tasks, chat and documents The **Agent** reads your team's work as well as your store's figures. In the Agent panel (the sparkle on the right edge of every page) you can ask about tasks and projects, chat channels and threads, documents and outreach. See for the Agent itself. It looks everything up as you. Two people asking the same question can get different answers, because each answer comes only from what the person asking is allowed to see. ## What can you ask about your work? [#what-can-you-ask-about-your-work] Anything you could find by opening the right pages yourself, in one question. Ask the way you would ask a colleague who has read everything. Questions that work well: * "What is late on the Spring launch project?" * "What did we agree with Portside Mills in #buying this week?" * "What does the returns policy document say about worn product?" * "Which outreach is assigned to me, and which is most urgent?" While it works, each step shows above the answer, such as "Looking through tasks", "Reading chat", "Reading documents" or "Reading outreach". How to follow those steps is covered in . ## What does it read about tasks and projects? [#what-does-it-read-about-tasks-and-projects] Every task in the workspace that is not archived, which is the same set everyone sees on the pages. It can list your own tasks, a project's or a department's, and find a task by the words in its title. For one task it reads the title, status, department, project, assignee, due date, description, subtasks, blockers and the newest 20 comments. So "what is holding up Priya Raman's sample approval task?" gets a real answer rather than a status. Your workspace may use its own word for this; **Settings › Terminology** shows both. ## What does it read in chat? [#what-does-it-read-in-chat] The channels you can see and the threads inside them, starting with the most recent messages and going further back when the question needs it. Name the channel in your question, such as "summarise what #buying said about the autumn polo order". With a channel or thread open, it reads that one with your question, so "what did we decide here?" is enough. See for channels and threads. ## What does it read in documents? [#what-does-it-read-in-documents] Your workspace's documents. It finds a document by the words in its title, lists the documents in a folder, and reads the one you name. Name the document by its title, or open it first. With a document open, "what does this say about lead times?" needs no title at all. See . ## What does it read about outreach? [#what-does-it-read-about-outreach] Open outreach items, either your own or the whole workspace's, with what each one is about, its stage, how urgent it is, who has it and the customer's name. For one item it also reads the newest 20 notes. With **View customers** on **Settings › Permissions**, it also reads the customer's lifetime spend. See for outreach itself. It never gives a customer's email address or phone number. Open the outreach item or the customer to find those, per . ## How does the page you have open come along? [#how-does-the-page-you-have-open-come-along] When you have a task, project, document, channel, thread, outreach item or customer open, the Agent reads it with your question. A chip reading **Using:** and the record's name shows which one, and its **X** stops using it. A customer you have open comes with their name, lifetime spend and number of orders, when you hold **View customers**. The full list of records the chip covers is in . ## Does reading your work need the Brain module? [#does-reading-your-work-need-the-brain-module] No. The Agent reads tasks, chat, documents and outreach whether the **Brain** module is on or off. The module decides only whether it can also suggest changes. It is on unless an admin switched it off on **Settings › Modules**, per . ## Can it turn what it read into a change? [#can-it-turn-what-it-read-into-a-change] Yes, with the Brain module on. Ask "make a task for Priya Raman to follow up on what #buying agreed" and it answers with a review card. Nothing changes until someone approves it, as covers. Ask for changes in a conversation that has not read a private channel. Once a conversation has read one, the Agent still answers but says it cannot suggest a change there. Press **New chat** and ask for the change in the new conversation, so nothing from a private channel reaches people outside it. ## Who can ask about their work? [#who-can-ask-about-their-work] Everyone who holds **Use the agent** on **Settings › Permissions**, which every role holds by default. See . Beyond that, what it reads follows what you can already open: the channels you can see, the workspace's tasks, documents and outreach, and customer spend when you hold **View customers**.