Use case · Research

Answer a market question with sources, not guesses

To do market research with AI agents, ask your Brainwrite Chief of Staff a specific question and let it split the work. It delegates bounded tasks to research bots: one reads public sources on the category, one reviews competitors, one reads customer discussions. Each returns evidence with links. The Chief writes one summary that separates what the sources say from what is inferred.

Download for Mac See pricing

Market research

I am deciding whether to build a bookkeeping add-on for dental clinics in the UK. Find out how clinics handle bookkeeping today, which tools they use, what those cost, and what owners complain about. Cite every source.

Your Chief of Staff put a team on it

Nova· Chief of Staff
Atlas· Desk researcher
Sage· Customer voice researcher

Summary: 3 findings with sources, 2 open questions the public sources could not answer
Tools: 4 bookkeeping products named in clinic discussions, with public pricing pages linked and dated
Customer voice: 12 public threads; the most repeated complaint is reconciling card payments
Inferred, not sourced: clinics with several sites likely outsource bookkeeping

DoneThe question and the decision it feeds waits for you. Nothing was sent.

An illustrated example. Your team works on your own files and apps.

Two colleagues review a laptop at a shared table while illustrated AI agents Nova, Atlas and Dash sit with them as teammates.

The job

Doing it yourself, and handing it off.

Doing it yourself

Research means twenty browser tabs: industry reports, competitor pricing pages, review sites, forum threads, public filings. You copy quotes into a doc, lose track of which tab a number came from, and run out of time before the synthesis. A single chatbot answer is faster but often does not show its sources, so you cannot tell which parts are grounded and which were filled in.

What the team does

Split the question into bounded research tasks
Read public sources and keep a link and date for every fact
Collect how buyers describe the problem in public discussions
Separate what sources say from what is inferred

What waits for you

The question and the decision it feeds
Which sources to trust when they disagree
Anything that involves paying for a report or signing up to a site

See it in the app

What it looks like in Brainwrite.

A real screenshot from a demo workspace.

Chief of Staff Juno proposing a four-bot Market Research team (Mira, Ledger, Sift, Quill) with each bot's role, model and Apply setup / Cancel.

The team

Who's on it.

Your Chief of Staff picks the specialists the job needs. Each has its own role, instructions, model, and app access.

NovaIllustrated example
Chief of Staff
Breaks the question into parts, delegates each one, checks what comes back, and writes the final summary.
AtlasIllustrated example
Desk researcher
Reads public sources in its built-in browser, keeps the link and date for every fact, and flags weak sources.
SageIllustrated example
Customer voice researcher
Reads public discussions, such as Reddit threads, for how buyers describe the problem in their own words.

How it runs

A run, step by step.

01
Ask a sharp question
Tell the Chief what decision the research supports, the market, and how deep to go.
02
Delegate
The Chief hands each part to a specialist. Each delegated task keeps its own activity and returns its result to your conversation.
03
Watch or step in
Open a bot's built-in browser tab to see what it is reading, or take over when a page needs you.
04
Get the summary
The Chief combines the findings into one summary with every source listed, saved in its working folder if you ask for a file.

What you'll connect

Apps this job uses.

FAQ

Questions about Market research

How is this different from asking one chatbot?

The Chief splits the question into bounded parts and gives each to a specialist with its own instructions, tools, and model. Each delegated task keeps its own activity, you can open the built-in browser to watch what a bot is reading, and every claim in the final summary comes with a link you can check.

Can I watch what the research bot is browsing?

Yes. Each bot can use a built-in browser tab that you can watch and take over. If a page needs a login or a decision, you can step in yourself. Watching the live preview does not give the bot any extra permissions.

Where do the results end up?

In the conversation, and in the bot's working folder on your Mac if you ask for a file. You can also have it write to Google Docs or Notion through connected apps. Bot memory can keep preferences such as trusted sources for the next project, and you can read or edit it at any time.

Built on