Three ways to chat

docAnalyzer has three chat modes. They share the same engine; what differs is what each one sees and what tools it can use. Knowing which mode does what makes the rest of the product much easier to navigate.

Ask docAnalyzer: the entry point

Ask docAnalyzer is the workspace-aware chat. It knows every document, note, and label you've uploaded, and it knows the product itself (plans, formats, limits, FAQ).

Use Ask when:

• You don't yet know which documents are relevant.
• You want a plan ("what should I look at first?", "how should I tackle this?").
• You're asking an orientation question ("what file types do you support?", "how do credits work?").
• Your request spans several different scopes and needs to be decomposed.

Ask doesn't usually answer with deep citations of its own. Instead, it spawns a Focus on the right documents, opens Co-work on a note, or suggests a workflow when the task is batch-shaped.

Focus: deep analysis on a dataset

Focus is a chat narrowed to a specific set of sources, your dataset. A dataset can be one document, many documents, a label (which expands to all its members), notes, or any mix.

Use Focus when:

• You know which sources to dig into.
• You want exact citations back to pages, sections, or cells.
• You want to produce a downloadable artifact (a report, spreadsheet, chart, or diagram) grounded in those sources.

Single-document, multi-document, and label-scoped chats are all the same Focus. There's no separate "per-document chat" feature; you open a Focus with a dataset of size 1, size N, or size N-after-label-expansion.

Co-work: collaborative note editing

Co-work opens a Note in the canvas editor and edits it alongside you. As you converse, the AI applies edits directly to the note: drafting sections, tightening prose, restructuring, rewriting in a different tone.

Use Co-work when:

• You've saved a Focus answer as a note and want to polish it.
• You've drafted something and want the AI to refine it in place.
• You're working on a long document where every turn produces visible changes you can accept or revise.

Co-work's context is the note itself, not a dataset of sources.

The loop

The three modes are designed to hand work back and forth:

1. Ask helps you decide what to look at.
2. Focus produces analysis grounded in the right sources.
3. You save the answer as a Note.
4. Co-work refines the note with you.
5. The refined note feeds back into a new Focus dataset, into a workflow, or out as a downloadable file.

Next: When to use which chat is the decision guide.