How is sensitive data protected?

All documents are encrypted with AES-256-GCM at rest and TLS 1.2+ in transit. A GDPR anonymization layer removes PII before any AI processing. Hosted on Vercel + Supabase (both SOC 2 Type II); Decision Intel’s own product-level audit is targeted for Q4 2026.

How long does integration take?

Less than 30 minutes. Upload documents directly, connect via OAuth for Slack, or use our REST API for bulk processing.

How does outcome tracking work?

Outcomes are detected automatically from follow-up documents, Slack messages, and web intelligence. You confirm with one click. Each reported outcome makes your future analyses more accurate.

How is Decision Intel different from ChatGPT or a general AI assistant?

ChatGPT gives one opinion from one model: ungoverned, untraceable, unaudited. Decision Intel is the reasoning audit platform. Most tools audit your data; we audit your reasoning — and catch the fatal blind spots in strategic memos before the committee does. It measures the noise in your reasoning the same way Kahneman did in the insurance underwriter study, simulates an AI boardroom of CEO, CFO, and board personas, runs what-if interventions against a 143-case public reference library, and compounds every confirmed outcome back into a calibrated Decision Quality Index your audit committee can defend. Not a chatbot; a reasoning audit, checkable from memo to outcome.

All workflows

For any high-stakes decision, before commitment

Decision pre-mortem

For any high-stakes decision, before commitment

Run a structured pre-mortem before the decision is final — surface failure modes the room cannot see from the inside.

Situation

You've reached the decision point. The room is converging on a recommendation.

Complication

Klein & Mitchell 1995 + Mitchell-Russo-Pennington 1989 found that asking the room "what could go wrong?" produces conditional-voice hedging. Asking them to describe the disaster in past tense produces 25-30% more, higher-quality failure-cause insights.

Question

What does the failure look like a year from now, in past tense — and what is the warning sign everyone missed?

Answer

Decision Intel runs the prospective-hindsight pre-mortem on the decision memo: projects one year forward, frames the outcome as a total disaster, generates the history of how it failed. Identifies the warning sign in the original memo before the decision is committed.

How it works

Upload the document

Paste the memo or drop the PDF / DOCX / spreadsheet. The 12-node R²F pipeline runs in ~60 seconds.

Review the audit

Decision Quality Index score, 22-bias R²F detection, named compound failure patterns, and verbatim evidence quotes for every finding.

Read the hardening questions

Each finding ships with an audit-committee-ready hardening question — the specific question a procurement reader would ask if they saw this memo cold.

Generate the prospective-hindsight narrative

The pre-mortem agent projects forward one year, frames the outcome as a total disaster, and writes the past-tense history of how it failed. Anchors the narrative on real bias patterns detected in the memo.

Lock the priors record

Before commitment, capture the falsifiable 90-day operational proxies the decision rests on. The Outcome Gate auto-fires at 90 days to score them. This is the calibration loop closure that compounds quarter after quarter.

Why this matters

A pre-mortem is the single highest-leverage discipline in strategic decision-making per the Kahneman literature. The audit operationalises it: every decision gets the pre-mortem treatment, every failure mode gets a falsifiable proxy, every outcome gets logged. The calibration loop is the moat, not the audit.

Related from the case library

Long-Term Capital Management

Long-Term Capital Management Collapse

Enron

Enron Accounting Fraud and Collapse

Lehman Brothers

Lehman Brothers Collapse

FAQ

What's the difference between pre-mortem and red-teaming?

Red teams scale poorly: they degrade into performative "list-the-risks" exercises because the political cost of dissent is unsustainable. The pre-mortem agent absorbs the antagonist role — the human user becomes the facilitator, not the antagonist. The system fires the dissent that the room cannot fire from the inside.

How do I structure the priors record?

The priors panel asks for: conviction level (0-100), rationale (≤500 chars), kill criteria (the falsifiable signals that would change your mind), and 1-3 micro-predictions with 90-day horizons. The Outcome Gate auto-fires at 90 days to score them with Brier loss.

Can I run a pre-mortem on a decision I already made?

Yes — the retroactive audit mode handles closed historical decisions. Useful for calibration: how well-calibrated were the priors you actually wrote? Per-org Brier score sharpens as you log more closed outcomes against original priors.

Does this work for non-strategic decisions (hiring, vendor selection)?

The pre-mortem detector is workflow-agnostic — it fires on any memo with a recommendation + supporting evidence + projected outcome. Hiring memos, vendor RFPs, product-bet memos, market-entry plans all work. The platform is positioned for strategic decisions but the underlying audit applies broadly.

Run a pre-mortem

Other workflows

Strategic memo audit

Audit a strategic memo for cognitive bias and forecast errors in 60 seconds — before the board sees it.

IC memo pre-vote stress test

Stress-test an investment committee memo before the vote — catch Synergy Mirage, Winner's Curse, and Conglomerate Fallacy.

Board deck pre-presentation audit

Audit a board deck before the presentation — predict the toughest questions the room will ask.

Fund investment thesis audit

Audit a fund investment thesis before the LP raise — catch Narrative Coherence, Reference-Class Blindness, and Optimism Trap.

M&A bias audit

Audit an acquisition thesis for the three named M&A failure patterns: Synergy Mirage, Conglomerate Fallacy, Winner's Curse.

Run the audit on a real memo.

Paste a strategic memo and see the audit run end-to-end — free, no card.