Put your thesis in front of a research team that disagrees.

How a meeting runs

1. 1You ask the question — “Is the margin recovery durable?”, “What would break this thesis?” — and choose the format and the panel.
2. 2InvestIQ builds the evidence pack: scores, financials, segments, estimates and consensus, earnings, peers, price — and the company’s Deep Research report when there is one. Each item gets a reference, like [E3].
3. 3Each analyst writes a private memo before seeing the others, so the first opinions are independent.
4. 4A moderator chairs the discussion: 20 to 40 messages, choosing who speaks next and on which point. When the pack lacks a figure, an analyst can ask the research desk, which looks in market data and SEC filings first and on the web last. You can interject while they talk.
5. 5Each analyst votes in private, then the verdict is written.
6. 6Afterwards, you can put follow-up questions to any analyst.

Choose the panel

Roles: Quant, Fundamental / Value, Growth, Macro, Technical, Risk manager, Short seller, Bull advocate, Credit analyst, Regulation & ESG, Sector specialist, Devil’s advocate, Behavioral finance — or a role you describe yourself.

Temperaments: balanced, skeptical, cautious, assertive or contrarian.

Formats: a debate; a red team that attacks your thesis until it holds or breaks; a pre-mortem (“a year later it failed — why?”); a Delphi round, where anonymous estimates converge round by round.

Different AIs: each seat can run on a different AI, so the analysts do not all share the same blind spots. If an AI is unavailable, another takes its seat and the debate says so.

Every figure is checked

Analysts must cite the evidence they use. For every sentence that states a figure, code checks that the figure is in the cited item — allowing for rounding, percentages and units, in English or French. No AI is involved in that check: a checker that could make things up would defeat the point.

• Verified: the figure is in the item cited.
• Unverified: an item is cited, but it does not hold that figure.
• Unsourced: a figure with no citation at all.

The result is shown under each message and, per analyst, in the verdict. The moderator computes no figure and adds no argument of its own.

What the verdict contains

• The team’s answer to your question, and how strong the consensus is — strong, moderate, split or none.
• What everyone accepted, and where the panel split, analyst by analyst.
• The cruxes: what the answer hinges on.
• For each analyst, the observable condition that would reverse their view.
• Blind spots — what nobody examined or what evidence was missing — and the questions worth a new meeting.

The verdict never says buy, sell or hold, and never sizes a position.

What the War Room does not do

• Decide for you: it lays out the arguments and their evidence; you judge.
• Give a recommendation, a price target or a position size.
• Let an analyst invent a figure unnoticed: unsourced and unverified figures are flagged.
• Predict the price: the analysts argue about the business and what the price implies.

On which plan

Details on the pricing page.