Skip to content
SynthBoard

Research and analysis

Research that ends in a call.
Not a pile of tabs.

A market to size, a claim to check, a forecast to build. On a board each expert researches on its own and says what it read before reacting to the others, so the disagreement is about evidence. Ask for the analysis and it comes back as the whole piece.

Start freeSee how it works

Free to start. No card.

Example. You write: “For my investment memo: is the “remote workers are 13% more productive” claim still a fair citation, how big is AI tutoring in Spain, and give me three adoption scenarios to 2030.” She opens 3 rooms: Is the 13% claim still fair? (The Scientist, up to 65 credits); AI tutoring in Spain: market size (The Analyst, up to 65 credits); Three adoption scenarios to 2030 (The Futurist, up to 65 credits). Nothing runs until you tap.

Examples

The market, the claim,
the teardown.

The ask as a researcher would type it, how she read it, who sat, and what came back. Examples written for this page, not recorded runs.

Which market size?

Work · Board of 3

I’m writing an investment memo on AI tutoring in Spain. Top-down reports say €900M. My bottom-up — 1.2M secondary students at €15 a month — says €216M. Which number goes in the memo?

Chief of staff

Two methods, one memo — worth a board. I’d seat The Analyst, The Skeptic and The Investor.

  • The AnalystData Expert
  • The SkepticRisk Assessor
  • The InvestorVC Partner

The board’s call

Lead with the bottom-up €216M as the market you can serve; show €900M only as the ceiling, with its source.

1.2M students at €15 a month for a year is €216M — and that assumes every one pays. Investors trust a number they can rebuild; the €900M rolls in adult learning you aren’t selling to.

Confidence76 · High

The Investor disagreed

“€216M at full adoption is a small market. Say what share you’d really win, or the memo argues against itself.”

First step Add 5% and 15% adoption rows to the sizingtomorrow

Example · Board of 3 · up to 135 credits

Does the claim hold?

Study · One expert

My thesis chapter cites “remote workers are 13% more productive”. My advisor says it’s outdated. Is it?

Chief of staff

The Scientist takes this — what the study measured, and what later work found.

  • The ScientistResearch Expert

Asked first

Is your chapter about all remote work, or fully remote versus hybrid?

All remote workFully remoteHybrid

The call · The Scientist

Keep the study, narrow the claim: it measured call-centre staff working from home, not remote work in general.

Cite it for what it found, then add the later studies on hybrid and fully remote work, which don’t all agree. That’s a stronger paragraph than either claim alone.

Example · One expert · up to 65 credits

The pricing teardown

Work · One expert

Write a two-page teardown of the market leader’s pricing for my deck: the tiers, what’s gated, and where a newcomer could undercut them.

Chief of staff

The Strategist writes it from their public pages, and says what it read.

  • The StrategistCompetitive Expert

The piece · The Strategist

“Three tiers, one wedge: the leader gates reporting behind its middle tier, which is exactly where schools…”

Two pages: the tiers in a table, what each gates, the two price points customers complain about in public reviews, and one wedge for a newcomer — with every source listed.

Example · One expert · up to 65 credits

What she handles

The evidence, kept.
The analysis, written.

What you’ve found stays on file with its source; the analysis comes back whole; the forecast comes with its range.

The analysis, written out.

Ask for the sizing, the comparison or the forecast and it comes back as the whole piece, sources named.

The question behind the question.

When the answer depends on what the work is for, the expert asks first. Being asked is never charged.

What you’ve found, kept.

Datasets, deadlines, the numbers you settled on. Each fact keeps where it came from.

A forecast with its range.

Three scenarios, the one to plan on, what would move it — and who thinks the middle case is still too kind.

The experts

Evidence first.
Then the argument.

The Scientist asks what result would prove you wrong, The Data Scientist asks whether the sample can carry the claim, The Analyst puts three scenarios on one page. On a board, each researches alone and says what it read.

Pick an expert

Hover, tap or tab throughSwipe, then tap one
  • GPT
  • Claude
  • Gemini
  • Grok

Seats are spread across model families: GPT, Claude, Gemini, Grok and more.

Questions

Fair questions.
Straight answers.

What people ask before they bring her this kind of call.

Or see pricing · every use case

Does it say where things came from?

Experts say what they read, and a board’s experts research on their own before reacting. Check anything that matters before you publish it — they’re AI and can be wrong.

Can it work with my own data?

Yes. Paste it, attach it or photograph it, or connect Google Sheets, Notion or Confluence. She keeps what matters on file, with where it came from.

Can it write in the register I need?

Ask for the piece in the shape and voice you need — a literature summary, a two-page teardown, a forecast with scenarios. It’s a draft to work from and check, not a finished source.

What does it cost?

Free to start — 150 credits every month + 250 to start. Talking to her is free within a daily allowance. A consult’s card says up to 65 credits and a three-seat board’s says up to 135 credits; nothing runs until you tap, and a board never bills above its card.

Close to this

For you

Consultants & advisors

Your client hears one answer. Hear the objection before they do.

Decision

Market entry

A new market is a two-year bet. Argue it before you hire for it.

Topic

Product & technology

Build, buy, or wait. Argued by experts who disagree.

Bring the question your research has to answer.
Start free, with one message.

Which market size goes in my memo?Is this study still a fair citation?Three adoption scenarios to 2030

It waits in her box after you sign up. Nothing is sent until you press send.