User guide
Documents
You want to write a letter, a handout or a report, or you have a Word file to change. This is the word processor, and how the agent writes in it.
The document program is a word processor for letters, notices, minutes and reports: headings, lists, tables, links and pictures in one flowing page, with Word files in and out and a PDF when you need one. It opens as a tab in the chat’s browser panel, so the agent can write into the document you have open while you watch.
This article is for anyone who writes: how to start a document, what the formatting bar does, how Word files come in and go out, how to print, and what to ask the agent for.
Making a document
Choose File ▸ New ▸ Document. The Documents list opens in a new tab of the browser panel with the name box ready: type a name and press New document. File ▸ Open ▸ Document…, or View ▸ Docs ▸ All Docs, opens the same list. Where documents are kept, and how saving and versions work, is the same for every program and is in The office.
The bar across the top shows the document’s name, whether it is saved, Save, and ⋯ (More), which holds Versions…, Rename…, Export…, Print or save as PDF…, All documents and Delete….
The formatting bar
Below that is the formatting bar, from left to right:
- the text style: Paragraph, Heading 1, Heading 2, Heading 3 or Code block;
- B, I and U for bold (Ctrl+B), italic (Ctrl+I) and underline (Ctrl+U);
- a bulleted list (Ctrl+Shift+8) and a numbered list (Ctrl+Shift+7);
- Link, which asks for a Link address and Apply, and Image;
- undo (Ctrl+Z) and redo (Ctrl+Shift+Z);
- More ▾, with Strikethrough, Inline code, Quote, Horizontal line, Align left, Align centre, Align right, Justify and Insert table.
Tables
More ▸ Insert table adds a table. While you are in a table, a Table bar adds a row above or below (↑ Row, ↓ Row), a column to the left or right (← Col, → Col), and deletes the row, the column or the whole table (✕ Row, ✕ Col, ✕ Table).
Pictures
Image picks a picture from your computer; you can also paste one or drop one in. A picture can be PNG, JPEG, GIF or WebP, up to 10 MB; an SVG file is refused. Pictures are kept inside the document, so a document with several large ones grows large itself.
Word files in and out
Import…, beside New document on the list, brings in a Word .docx file as a new document, up to 20 MB. Other formats are refused. What could not come in is listed in a note on the new document — “Imported from …. Not brought in: …” — which can name page headers and footers, footnotes and endnotes, comments, text boxes, fields (such as page numbers or a table of contents), embedded objects, equations, page breaks, macros and pictures in a format the document cannot show. Tracked changes are brought in as accepted.
⋯ ▸ Export… saves the document as a .docx file, asking where, with the document’s name already filled in.
Printing and PDF
⋯ ▸ Print or save as PDF… lays the document out in pages. Choose the Paper (Letter or A4; it starts at the one your computer’s region uses) and the Margins (1.5 cm, 2 cm, 2.5 cm, 0.75 in or 1 in). Then:
- Save as PDF… writes a PDF where you choose.
- Print… makes the PDF and opens it in your computer’s PDF viewer, which is where you print from.
- Back to the document returns to editing.
Asking the agent
Ask in your own words: write a welcome letter, add a section on parking after the introduction, change “Saturday” to “Sunday” everywhere, make the closing paragraph a quote, put the shift times in a table, tidy up what I’ve selected, save a Word copy. The agent works in the document open in the chat’s browser panel, opening it there first if it needs to.
- Reading: it can see the document’s headings, read the whole document or one section, and read what you have selected.
- Writing: it adds text with headings, lists, emphasis, links, quotes and tables — before a heading, right after it, after its whole section, at the end of the document, or in place of your selection (at the cursor when nothing is selected).
- Rewriting: it replaces a word or phrase everywhere, or in one section, and says how many it changed.
- Styling: it makes passages headings, paragraphs, lists or quotes; bold, italic, underlined, struck through or code; and aligned left, centre, right or justified.
- Tables and pictures: it builds a table from rows of values, and can place pictures held in one of your databases.
- A Word copy: it saves the document as a .docx file in your Downloads folder.
- Bringing one in: it can bring in a .docx file you attach to the chat, as a new document, and tell you what was left out.
Each request lands as one change. A note above the page says what the agent did — “The agent inserted content at the end.”, for instance — with Undo, which takes back that change.
What it asks
- Without asking: reading the document, its headings and your selection.
- Asked once: making a new document or bringing one in from a file — allowing that also lets the agent change it in this chat. For a document you already had, the first change asks, and its main answer is Allow changes to “…” in this chat; Allow once is behind the arrow beside it. Saving a Word copy counts as a change.
- Each time: nothing the agent does in a document is asked about every time, because it has no way to delete one. Deleting a document is yours, from ⋯ ▸ Delete….
Approvals and permissions explains the answers in full.
Good to know
- A document is one flowing page on screen. It has no page headers or footers, footnotes, page breaks, comments or tracked changes; pages appear when you print or make a PDF.
- Only you can make a PDF. The agent saves Word copies only.
- A document can be up to 32 MB. Pictures are what make one large; removing a large picture makes it smaller.
- The agent adds at most 20,000 characters in one step, so a long document is written in several.
- A document is kept as the application’s own file, not as a Word file. To hand one to someone, export it or save a PDF.
Back to all user-guide articles.