Product

How Decision Copilot works

You bring the decision. Independent models pressure-test the same facts. You get structured briefs — and, when you run a think tank, one merged Unified Brief you can discuss, attribute, and export.

The four steps

You bring the decision. Decision Copilot brings independent perspectives pressure-testing the same problem, then creates a Unified Brief with plans you can refine and act on.

Describe what's on the table
Lay out the decision as you see it — situation, what you know, what's still unclear. Then tell your think tank how to analyze it: compare options, gut-check a favorite, show the opposition case, surface risks, or widen the alternatives — so the analysis matches the kind of help you need.
Convene your think tank
Run a single AI model or several at once. Each one reads the same brief and produces its own structured analysis — so you're not relying on one voice for a complicated call.
Answer what the models need to know
Before finalizing, the models ask follow-up questions — missing context and details they can't safely assume without your clarification. Your answers sharpen the analysis.
Get a brief you can use
Each run produces a structured decision brief — risks, tradeoffs, stakeholder impacts, recommendation, next steps. Read it, edit it in place, and discuss it with the models. When you've run a think tank, merge into one Unified Brief — then challenge the synthesis or regenerate when you want the merge to change.

What's in a Unified Brief

One merged brief from your think tank — written blind by default so ideas beat brand names.

How it's built

Run a think tank and each model produces its own Decision Brief. You pick a synthesizer to merge those into one Unified Brief — a structured write-up, not a chat transcript.

What's in the document

Situation & summary

The merged read of what's on the table.

Risk, Reversibility, Stakeholders

What could go wrong, how hard it is to undo, and who bears the consequences.

Recommendation

One clear call from the think tank's best ideas.

Key considerations & next steps

Tradeoffs to keep visible and concrete next steps you can act on.

From chat

Spun off from your conversation with the models and folded into the same document.

Research

An optional “show your work” addendum. Ask a model to justify its reasoning and the supporting lookups and evidence it surfaces are saved here.

Variants

Alternate framings or what-if branches you spin off from chat and keep, so the brief can weigh them too.

Postures — how you frame the analysis

At intake you choose how models analyze the decision. Every brief still runs through the same three lenses; the posture sets the emphasis:

Compare options openly

Balanced paths — no preferred direction assumed.

Challenge my leaning

Pressure-test your plan; downsides and blind spots.

Show me the opposition

Steelman the strongest case against your lean.

Risk-first

Downside scan — risks and hidden assumptions first.

Widen the option set

Alternatives and adjacent factors you may not have named.

Prefer a different shape? Use freeform and let the model choose its own structure.

Chat is where the work happens

A brief isn't a static document — it's a conversation. Chat with any model, or the synthesizer, to press on the reasoning, challenge a call, and reshape the brief in place. It's also where you spin off the research and variants that fold back into the document.

  • Decision Briefs — chat in Discuss & clarify; edit sections in place; ask for new views.
  • Unified Brief — discuss with the synthesizer or any reviewer; ask why an idea was left out; regenerate when you want the written merge to change.
  • Contributions — see what each model contributed after the blind merge; export as PDF with an optional appendix.

Fact-check pass (optional)

A separate judge with web search can flag public factual errors and rewrite the draft — without changing the recommendation.

Start a decision →Try a demo →
← Back to homepage