Skip to main content
GBrain
Menu
Get Started

Research

The deep dive you need, when you need it.

What comes in, what it makes, where it lands

reads

Deep Person Brief

Research a candidate, podcast guest, or investor before you meet. A fully sourced brief from public sources and your team's notes, so you walk in already prepared.

Set up Deep Person Brief

reads

Fact Check

Verifies every checkable claim in your draft against live sources, corrects what is wrong, and returns a scored report before you publish.

Set up Fact Check

reads

Study Check

Traces a cited study back to its original paper, data, and replications, then returns a plain verdict on whether the claim holds up.

Set up Study Check

reads

Applied Reading

Reads a book, article, or case study through the lens of a problem you are facing and returns an applied playbook of moves, not a summary.

Set up Applied Reading

Gmail, Outlook, Google Drive

Gmail
Outlook
Google Drive
reads

Structured Data Pull

Pulls the fields you care about out of emails, documents, and web pages into one clean, deduplicated table, with every row linked to its source.

Set up Structured Data Pull

Gmail, Outlook

Gmail
Outlook
reads

Pitch Scorecard

Scores each inbound pitch against the bar you define and tells you which ones deserve your time, with the reason spelled out.

Set up Pitch Scorecard

Every email waits in Gmail or Outlook Drafts for a person to send. The five share one customer list.

Not waiting for Monday's digest. Answering the question on your desk right now.

Research, on your clock

Research abilities are the opposite shape from reporting: one-off, ad-hoc, triggered by a specific question rather than a recurring schedule. Someone on your team hits a decision point — “what have we actually learned about vendor X in the last eighteen months?” — and they want the answer by lunch, not next Monday.

These abilities are templates for that conversation: a described research task, a set of sources to pull from (your team’s documents and your team’s watched feeds, plus any public source you name), and a written output in the shape you need.

Research abilities share a common shape:

  • A question or task — what the team member is trying to decide or understand.
  • A scope — which sources, which time range, which people, which parcels.
  • An output — a dossier, a comparison table, a summary, whatever your team can act on.
  • A turnaround — usually minutes, not days.

Unlike reporting, research doesn’t run on a cron. It runs when a teammate opens the agent, describes the task, and hits go.

It can read the live web, not just simple pages

A lot of the sources that matter (city and county meeting portals, planning dashboards, permit lookups) only show their content after the page loads and runs. A basic fetch sees a blank shell. Your harness opens these the way a person’s browser would, waits for the page to fill in, and reads what actually appears, so a JavaScript-heavy government site is no longer a dead end. It can also grab a screenshot of a page as evidence when you want to see exactly what it saw.

Turn one on, then the rest

Each one sets itself up in a chat: it asks where your customers write, sends you to connect the accounts it needs, and runs. Turn on a second one and it finds the customer list the first one made.

Start with Deep Person Brief