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.

Which cognitive biases caused Celsius Network's collapse?

5 bias patterns are documented in this case, led by Overconfidence Bias. The full set: Overconfidence Bias, Planning Fallacy, Confirmation Bias, Framing Effect, Sunk Cost Fallacy. Each is mapped to Decision Intel's 22-bias R²F taxonomy with stable IDs and academic citations.

What is Decision Intel's reasoning audit?

Decision Intel runs a 12-node Recognition-Rigor Framework (R²F) pipeline on strategic memos, detecting cognitive biases from a 22-bias canonical taxonomy and producing a Decision Quality Index (DQI) score. The audit produces a hashed, tamper-evident Decision Provenance Record (DPR) suitable for audit-committee review.

All case studies

Catastrophic failureFinancial Services2022

Celsius Network

Celsius Network Bankruptcy

Estimated impact: $4.7B in customer deposits frozen

F14

Decision Quality Index (simulated)

Critical · 5 biases detected, 2 toxic combinations

Crypto lending platform promised 18%+ yields to depositors while deploying funds in high-risk DeFi strategies and proprietary trading. When crypto markets declined, a $1.2B balance sheet hole was revealed.

Decision context

CEO Alex Mashinsky publicly assured depositors funds were safe while internally the risk team documented growing losses. Yield promises were maintained even as returns from lending declined, requiring increasingly risky strategies to cover the spread.

Decision anatomy

Monetary stakes

Very high

Dissent suppressed

Yes

Time pressure

Present

Unanimous consensus

Yes

Decision participants

Dissent encouraged

External advisors

Iterative process

Red = risk factor present · Green = protective factor present

Decision Intel Platform Analysis

Source: Public Statement · 2022-03

The analysis below was produced from the pre-decision document only. No hindsight. This is what the platform would have surfaced if it had been running in 2022-03.

Public Statement · 2022-03

“Celsius Network's 2021 marketing materials promised depositors up to 18% annualized yield on stablecoin deposits — in a market where risk-free rates were near zero. CEO Alex Mashinsky hosted weekly AMAs repeatedly stating 'banks are not your friends' and that Celsius funds were safer than traditional banking. Internally, Celsius deployed deposits into DeFi protocols (Anchor 20%, Aave, Compound) and proprietary trading at Celsius KeyFi subsidiary. Internal risk reports flagged a growing mismatch between yield promised to depositors and yield generated by assets, plugged by CEL token appreciation dependencies.”

Source: Celsius Network Chapter 11 filings (SDNY Case 22-10964); Alex Mashinsky DOJ Indictment (July 2023); NYAG complaint

Red flags detectable at decision time

18% yield on stablecoins in a zero-rate environment — structurally unsustainable without deposit-funded losses
CEO public messaging inconsistent with internal risk-team assessments
Deposit/yield-source asset duration mismatch growing through 2021-2022
Proprietary trading losses (Celsius KeyFi) not disclosed to depositors
CEL token price used as solvency mechanism — reflexive dependency between platform and token value

Cognitive biases the platform would have flagged

Overconfidence BiasFraming EffectBandwagon EffectConfirmation Bias

Hypothetical analysis

DI would flag Celsius as the canonical framing-effect deception compounded by bandwagon effect. The 'banks are not your friends' framing converted what was functionally an unsecured loan to a crypto operator into a moral positioning statement — disarming normal depositor-risk scrutiny. A bias-adjusted depositor-facing communication would have been required to disclose the sources of the 18% yield, the asset-liability duration profile, and the CEL-token dependency. The bandwagon effect on the depositor side — 'millions of users are doing this' — substituted social proof for independent risk assessment.

Biases present in the decision

★Overconfidence Bias

critical

Planning Fallacy

medium

Confirmation Bias

high

Framing Effect

medium

Sunk Cost Fallacy

high

★ Primary driver · Severity estimated from bias type and decision outcome

3D Decision Knowledge Graph

Celsius Network · 5 biases, 2 toxic combinations

Drag to rotate · Scroll to zoom · Click to explore

Decision

Critical

High

Medium

Toxic

Bias Intensity Profile

Severity × confidence across 5 detected biases

Toxic combinations

!Optimism Trap

Overconfidence and availability bias create a systematically positive outlook that ignores base rates.

!Sunk Ship

Commitment to a failing course escalates to justify sunk costs, even as evidence mounts.

Reference class base rates

Across all 143 curated case studies in our library:

89%

base failure rate

-0%

partial / mixed

11%

base success rate

■ Failure■ Partial■ Success

Lessons learned

Yield promises create structural incentives for increasing risk
Customer-facing assurances must be independently verified against internal data
DeFi counterparty risk chains can create hidden systemic exposure

Source: Celsius Network bankruptcy filing, Chapter 11, SDNY 2022; DOJ Indictment 2023 (SEC Filing)

Pilot slots available

We caught these patterns in Celsius Network's own record — before the outcome.

See the full hindsight-stripped auditwe ran — no login, no card. Then run the same 60-second audit on your own next memo.

See a live audit · no login Audit your own memo

Or leave your email, we'll run a strategic memo of your choosing and send the readout within a business day.

Audit your own decision

Workflows that fire on decisions like Celsius Network’s

The same Recognition-Rigor Framework that documents this case audits memos in the same shape — before the outcome forces the lesson.

For CSOs and Heads of Strategic Planning

Strategic memo audit

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

Audit a strategic memo

For VC / PE General Partners

Fund investment thesis audit

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

Audit a fund thesis

For any high-stakes decision, before commitment

Decision pre-mortem

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

Run a pre-mortem

See all 6 workflows for the full audit-tool surface.

Related cases in Financial Services

Partial success2018

Access Bank PLC

Access Bank Acquisition of Diamond Bank

Catastrophic failure1998

Long-Term Capital Management

Long-Term Capital Management Collapse

Catastrophic failure2008

American International Group

AIG Credit Default Swap Crisis

Catastrophic failure2011

MF Global

MF Global Bankruptcy