> ## Documentation Index
> Fetch the complete documentation index at: https://docs.merciv.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Modes and output formats

> Choose how deeply the assistant works and what kind of output it produces.

Two pickers on the left side of the composer control how the assistant approaches your question: the **mode** sets how deeply it works, and the **output format** sets what it produces.

<Note>
  Some modes and the output format picker are available on select plans. Contact your Merciv representative if you don't see an option described here.
</Note>

## Modes

| Mode         | What it does                                                                                                               | When to use it                                                                                                                      |
| ------------ | -------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------- |
| **Work**     | The default. The assistant works through your question step by step — searching, reading, and analyzing before it answers. | Most questions, especially ones that need evidence from multiple sources.                                                           |
| **Quick**    | Fast answers with less digging.                                                                                            | Simple lookups and follow-ups where speed matters more than depth.                                                                  |
| **Research** | Long-running deep research that produces a full formatted report with a cover illustration.                                | Broad questions that deserve a thorough, citable document — competitive landscapes, category reviews, trend analyses.               |
| **Personas** | The assistant answers as one of your organization's configured personas.                                                   | Testing how a message, concept, or product lands with a specific audience. Only appears when your organization has personas set up. |

Research runs take longer than a normal answer — processing happens in the background and can take several minutes. The finished report appears as a report card in the message, sometimes a few seconds after the answer text.

## Output formats

The output format picker tells the assistant what to produce:

* **Auto** — the default. The assistant decides what fits the question best.
* **Report** — a written document.
* **Dashboard** — a set of charts and metrics.
* **Slides** — a presentation deck.
* **Spreadsheet** — tabular data you can download and work with.
* **Visualization** — a single chart or graphic.
* **Canvas** — an interactive dashboard rendered directly in the response, with charts, tables, metric tiles, and drilldowns you can explore in place.

The format resets to **Auto** after each message, so set it again if you want the next answer in the same format.

<Note>
  Canvas is available on select plans. Contact your Merciv representative if you don't see it.
</Note>

Generated documents, spreadsheets, and slides appear as file cards in the response — see [Generated files](/chat/generated-files) for how to view, copy, and download them.

## Voice dictation

A microphone button in the composer dictates speech into the message box, in browsers that support speech recognition. If the browser blocks microphone access or doesn't hear anything, a "Voice input" toast explains what happened.

<Note>
  Dictation is available on select plans. Contact your Merciv representative if you don't see the microphone button.
</Note>