Decision IntelDecision Intel
Pricing
Sign InBook a callTry the Demo
Inside the engine

How we audit a strategic memo.

The audit fires the constraint while the deal sponsor is still drafting — not after the IC memo lands. Twelve specialized agents grounded in Kahneman, Klein, and Tetlock. The audit log carries the provenance, so the corp dev lead routes the flagged signal to the sponsor without spending personal political capital on the challenge. The methodology backbone (taxonomy, frameworks, regulatory map) lives further down the page for procurement readers.

This is a general-but-detailed walk-through of our methodology. It omits proprietary weights and prompts by design. Everything you see here is public-safe, citable, and reproducible against the academic record.

Try it on a memo See the taxonomyRead the research
12-node pipeline · under 60 seconds
What happens when you press audit.
01 · Preprocessing3 nodes
Redact · Structure · Contextualize
02 · Analysis (parallel)7 nodes
Seven specialized agents run at once
03 · Synthesis2 nodes
Reconcile · Score deterministically
Scroll for the full diagram →DQI · 0–100 · A–F
Anatomy of a call

Every Decision Intel call composes five rigor layers around one decision.

Knowledge Graph (what you've decided before), AI Boardroom (how each stakeholder will receive the memo), Reasoning Audit (what the bias + noise stack flags), What-if (the counterfactual impact), Outcome Loop (how the decision actually played out). Individually, any one of these is a feature. Composed, they are the call. The pipeline below is how we assemble them.
GraphBoardroomAuditWhat-ifOutcomeDIEVERY ANGLE · ONE CALL
The pipeline

Twelve specialized agents. Three zones.

Every memo passes through a sequential preprocessing chain, then a parallel fan-out of seven analysis agents that reason over the same shared context, then a two-step synthesis that reconciles the signals and computes a deterministic score. Click any node to see what it does.
PreprocessingAnalysis · 7 parallel agentsSynthesisOutputDQI · 0–100 · A–F
GDPR AnonymizerPII redacted before any LLM see…
Data StructurerParses sections, speakers, and…
Intelligence GathererExtracts topic, industry, and r…
Bias DetectiveDetects 22 cognitive biases wit…
VerificationFact-checks claims and maps com…
SimulationFive steering-committee persona…
Forgotten QuestionsSurfaces the questions the memo…
Noise JudgeThree decorrelated samples acro…
Deep AnalysisLinguistic, logical, and strate…
RPD RecognitionPattern-matches against a histo…
Meta JudgeReconciles the seven parallel s…
Risk ScorerComputes the final DQI — determ…
01 · Preprocessing02 · Analysis (parallel)03 · SynthesisEach zone runs in order. Inside Analysis, all seven agents run simultaneously against the same shared context.
Bias detection

Thirty-plus cognitive biases. Every detection citable.

Our taxonomy is published openly at /taxonomy (DI-B-001 onward, growing as the Kahneman-Klein paper-application sprint lands new detectors), extended with eleven strategy-specific biases drawn from Stanford VC and PE decision research. Every detection comes back with an excerpt, a severity, and a confidence score.
30+cognitive biases
22general (DI-B-001–022)
11strategy-specific
0detections without an excerpt
DI-B-001
Confirmation Bias

Selectively citing evidence that supports the dominant hypothesis and dismissing evidence that would reverse the call.

Kodak· 1975–2012See the case
DI-B-004
Groupthink

Suppressing dissent to maintain group harmony. The memo reads like unanimous consensus where there should be friction.

U.S. Government· 1961See the case
DI-B-005
Authority Bias

Senior-voice framing replaces independent judgment. The argument defers to position, not evidence.

Yale University· 1963See the case
DI-B-007
Overconfidence

Stated certainty far exceeds what the evidence supports. Confidence language without commensurate calibration.

LTCM· 1998See the case
DI-B-009
Planning Fallacy

Timelines and costs estimated bottom-up instead of against comparable reference classes. Understated by design.

NSW Government· 1957–1973See the case
DI-B-012
Status Quo Bias

Preference for the current path dressed up as strategic discipline. Inaction gets the benefit of every doubt.

Blockbuster· 2000–2010See the case
See all 22 biases with academic citations
Toxic combinations

Individual biases are features. Combinations are catastrophic.

Our 22×22 interaction matrix scores every bias pair against the others. Context amplifiers multiply the score when monetary stakes are high, dissent is absent, or time pressure is active. False-positive damping kicks in when a pattern gets flagged but the outcome succeeded. Over time, each organization calibrates its own weights from its own outcomes, which is why this section of the engine is the hardest to replicate.
Toxic network
How the biases combine
Inner ring: biases that participate in multiple toxic patterns. Edge color = pattern. Hover a pattern below to isolate its edges.
4Overconfidence Bias2Confirmation Bias2Groupthink2Anchoring BiasLoss AversionPlanning FallacySunk Cost FallacyStatus Quo BiasAuthority BiasIllusion of ValidityHalo EffectDecisionsat compound risk
11 biases · 10 named patterns · 10 toxic edgesHub numbers = pattern participation count.
M&A workflow coverage · by deal stage

Each pattern fires at the deal stage where it's most decision-fatal.

Document type drives overlay. A CIM gets the seller-halo filter and the strategic-adjacency audit; an IC memo gets the full toxic-combination pass plus the reference-class forecast; a synergy model fires the Synergy Mirage detector hardest. The same 12-node pipeline runs on each, with stage-appropriate prompts layered on top.
STAGE 1live

Sourcing

CIM · teaser · IM
Conglomerate Fallacy

Strategic Adjacency Audit. Zook core-vs-adjacency framework on the target's distance from the acquirer's core.

STAGE 2live

Diligence

QofE · model · DD pack
Synergy MirageReference Class Forecast

Synergy claims pattern-matched against BCG triad (mechanism / owner / 90-day milestone). Reference Class Forecast benchmarks projections against the 143-case base rate.

STAGE 3live

IC Review

IC memo · IC deck
Synergy MirageWinner's CurseYes Committee

Full toxic-combination pass. Boardroom Decision Twin simulates 5-persona IC vote. Pre-mortem fires prospective hindsight (Klein & Mitchell 1995).

STAGE 4live

Closing

LOI · final memo · term sheet
Winner's CurseSunk Ship

Deal Fever Check: scans final documents against original screening memo for escalation language and anchoring drift.

STAGE 5roadmap

Post-Merger Integration

Integration plan · Day-1 ops
Synergy Mirage (execution side)

Cultural divergence + IT-simplicity fallacy + talent-flight retention. Roadmap: deeper PMI overlays land once acquirer-side engineering capacity is in place.

Decision Quality Index

A FICO score for decisions. Zero to a hundred, A through F.

The final DQI is a weighted composite across six components. The weights are fixed and transparent; the scores inside each component are computed deterministically from the earlier pipeline outputs. Same inputs always produce the same DQI.
DQI v2.0.0 · six weighted components
How a memo becomes a score between 0 and 100.
Bias Load28%

Severity-weighted count of detected cognitive biases, normalized to document complexity.

Noise Level18%

Inter-judge variance from the three-judge noise panel. Low variance = stable reasoning.

Evidence Quality18%

Share of quantitative claims that verify against grounded search, plus source reliability.

Process Maturity13%

Was a prior submitted, outcomes tracked, dissent present, right committee size?

Compliance Risk13%

Inverse of the seven-framework regulatory exposure score from the Verification node.

Historical Alignment10%

Pattern match against 143 historical cases. Prior failure signatures drag the score down.

Grade scale
A85+

Board-ready. Strong reasoning across the stack.

B70–84

Mostly solid. Address the flagged biases before the vote.

C55–69

Mixed. Several reasoning gaps need explicit treatment.

D40–54

Weak. Rework required before the committee reviews.

F0–39

Critical. Reset the memo before re-circulating.

Sample score
Enron (Aug 2001)
38D

Groupthink + authority + off-balance-sheet masking.

Sample score
Apple iPhone (Jan 2007)
86A

Explicit risks, dissent tracked, reference class cited.

Sample score
WeWork S-1 (Aug 2019)
24F

Narrative fallacy, founder halo, undefined unit economics.

Three outputs on every audit · plain-language first

Validity. Outside View. Author Calibration. Three load-bearing outputs, three published papers.

Three signals fire on the audit before the DQI is finalised. Each one is named for what the buyer reads on the cover of every Decision Provenance Record. Plain language for the procurement reviewer, with the Recognition-Rigor Framework technical names + the source-paper citations carried as the supporting context. Each shifts the score, and each appears as a first-class strip on every audit so the methodology is verifiable against the source paper directly.
R²F · Detector 1

Validity

Technical name · Validity Classifier

Classifies the decision domain into high / medium / low / zero validity. High-validity domains have stable rules and rapid feedback (chess, weather forecasting one week out). Low-validity domains are noisy with delayed feedback (M&A, market entry, long-horizon strategy). Most strategic memos sit in the low / zero band.

What shifts in the DQI
Confidence-language is penalised harder in low-validity domains (the same rhetorical certainty that scores neutrally on a high-validity decision becomes an Illusion-of-Validity flag in low-validity contexts). Methodology v2.1.0.
Anchor: Kahneman & Klein, "Conditions for Intuitive Expertise: A Failure to Disagree" (American Psychologist, 2009) · first condition.
R²F · Detector 2

Outside View

Technical name · Reference Class Forecast

Pure-function similarity scoring against the 143-case reference-class corpus. Returns top-5 historical analogs + matched-class baseline failure rate + four-band predicted outcome (succeeds / mixed / struggles / fails / too-small-to-judge). Structurally novel decisions return "too small to judge" rather than a fabricated forecast: the cold-start posture is honest.

What shifts in the DQI
When the matched class shows a higher base-rate failure than the memo concedes, the audit flags Inside-View Dominance (DI-B-022) and the audit-committee-ready hardening question lands on the cover of the DPR.
Anchor: Kahneman & Lovallo, "Delusions of Success: How Optimism Undermines Executives' Decisions" (Harvard Business Review, July 2003).
R²F · Detector 3

Author Calibration

Technical name · Feedback Adequacy

Audits the second condition for trustworthy intuition. Has the decision-maker had enough closed-loop feedback in this specific domain for their experience to be calibrated? Verdict bands: adequate (≥10 closed outcomes in domain past 18 months) / sparse (3-9) / cold-start (<3) / unknown.

What shifts in the DQI
Cold-start posture: an audit by a domain-novice carries the same scrutiny rules as one with high closed-outcome history but the experience-based confidence claims are flagged for the reviewer rather than discounted silently.
Anchor: Kahneman & Klein (2009) · second condition (adequate opportunity to learn from rapid feedback).
Calibration anchor
The same R²F methodology applied retrospectively to 143 historical corporate decisions produces a mean Brier score of 0.258 · fair band, between CIA-analyst (0.23) and motivated-amateur (0.35) per Tetlock anchors.
See the calibration baseline
Counterfactual lift

What shifts when you remove the bias. Made visible.

Every flagged bias on a memo carries a reproducible lift weight: the exact number of DQI points that returns to the score if the bias is mitigated. Toggle the chips below on the actual WeWork S-1 audit and watch the gauge respond. Same memo, same lift, every time. The calculation is deterministic, not generative.
Real audit · public document
WeWork S-1
Public document · 2019 IPO prospectus · 3 biases flagged. Click any to mitigate it and watch the DQI lift.
Lift weights are calibrated against the same rubric as the live DQI calculation, so the same memo always returns the same number. The mitigated ceiling stays in the D-range because the underlying decision had structural failures beyond bias.
Decision Quality Index
24
FReject as drafted
24 as drafted · 24 if all flagged biases mitigated
Total lift available: +26 DQI points
Two parallel stability checks

Stable reasoning. Surviving dissent.

Two of the seven analysis agents run at the same time and answer complementary questions. Would a second read of the memo come to the same conclusion? And would it survive the real room?
Noise decomposition · three decorrelated samples
Is the reasoning stable under rewording?
Low noiseReliable
Mean
78/100
Std Dev
4

Three judges converge on the same score. The reasoning holds up — the memo says what it means.

High noiseUnstable
Mean
62/100
Std Dev
22

Same memo, same prompt, three different reads. Something is ambiguous — rewrite before the board sees it.

Boardroom simulation · five role-primed personas
Would this memo survive the room?
CF
Skeptical CFO· Capital discipline
Counter-case not stress-tested.
REVISE
CE
Ambitious CEO· Growth bias
Timing fits the cycle.
APPROVE
BC
Board Chair· Governance
Dissent absent from the memo.
REVISE
OP
Operator· Execution risk
Delivery path undefined.
REJECT
CO
Compliance Officer· Regulatory exposure
Frameworks handled.
APPROVE
Overall verdictMIXED
2approve2revise1reject3dissent tracked
The Decision Provenance Record

What comes out the end. Hashed, cited, and built to be defended.

Every audit produces a Decision Provenance Record: the artifact your team carries into the room when the decision is challenged. Each section below is a real part of the record we generate today; click any row to see what it contains, what regulatory provision it satisfies, and what an external reviewer can verify without leaving the document. Specimen records are publicly available so a procurement reviewer can read one before the conversation starts.
DECISION PROVENANCE RECORD · v1SHA-256 7b3a1f…d09
DPR section
Biases identified
Contains
  • Each detection: bias name + DI-B taxonomy ID
  • Verbatim excerpt from the source document
  • Severity score (low / medium / high / critical)
  • Citation back to the peer-reviewed paper that named the bias
Regulatory provision
EU AI Act Art. 13 — transparency to affected parties
What a reviewer verifies

Every bias name links back to a published academic source. A skeptical reviewer can chase any flag back to its original definition without leaving the document.

Read a real DPR (WeWork S-1, PDF)Pan-African shape (Dangote, PDF)
Closing the loop

Every decision becomes signal for the next one.

An audit is only half the work. After a decision gets made, Decision Intel listens for the outcome through the tools your team already uses (Slack threads, Drive folders, inbox replies, public announcements) and writes what actually happened back into your Decision Knowledge Graph. Your team's calibration improves quarter after quarter, on your own data.
After the audit · Outcome loop
Every decision closes its own loop — automatically.
detecting
Earlier memo · decision at rest
auditedscoredlogged
Loop closed
Signal convergence0/4
Slack
listening to threads
Drive
watching folders
Web
scanning public signals
Email
monitoring replies
Awaiting convergence
Signals still arriving…
Calibration pending·Once the outcome is detected, the loop writes back to the Decision Knowledge Graph.
Four detection channels. Zero manual logging. Every outcome compounds the baseline.See the full loop
Academic foundation

Standing on shoulders.

None of this methodology is invented in a vacuum. Every node in the pipeline cites a specific academic lineage, and every detected bias on your memo links back to the peer-reviewed paper that first named it.
DK
Daniel Kahneman, Olivier Sibony, Cass Sunstein
2021
Noise: A Flaw in Human Judgment

Ensemble sampling across three Kahneman-side nodes plus inter-judge variance scoring inside the Noise Judge node come directly from this framework.

PowersNoise Judge
GK
Gary Klein
1998
Sources of Power: How People Make Decisions

Recognition-Primed Decision theory grounds the RPD Recognition node: pattern matching against a labeled historical library.

PowersRPD Recognition
KT
Daniel Kahneman & Amos Tversky
1974 / 1979
Prospect Theory & Judgment under Uncertainty

The foundational taxonomy for framing, loss aversion, anchoring, and availability biases detected by the Bias Detective.

PowersBias Detective
PT
Philip Tetlock
2005 / 2015
Superforecasting & Expert Political Judgment

Calibration methodology for the outcome flywheel and the Forgotten Questions node: what are you not asking?

PowersForgotten Questions
AD
Annie Duke
2018
Thinking in Bets

Probabilistic decision framing behind the blind-prior capture and per-analysis confidence model.

PowersBlind priors
IS
Ilya Strebulaev
ongoing
Stanford VC Initiative · Corporate Decision Research

Source for the 11 strategy-specific biases (entry-price anchor, thesis confirmation, winner’s curse, management halo).

Powers11 strategy biases
RD
Ray Dalio
2021
Principles for Dealing with the Changing World Order

18 rise-and-fall determinants (debt cycle, currency cycle, reserve-currency status, governance, infrastructure) feed the Structural Assumptions audit, the macro-layer pass that sits beside cognitive-bias detection, asking what the plan is betting on about the world.

PowersStructural Assumptions
Security · privacy · compliance

The same standard we audit your memos with, we hold ourselves to.

  • GDPR anonymizer is the first node, not an afterthought. PII is redacted before any analysis LLM sees the document. If anonymization fails, the pipeline short-circuits to the risk scorer rather than transmitting raw content.
  • AES-256-GCM document encryption at rest; per-record keys; rotating encryption envelopes.
  • Seven-framework compliance mapping built into the Verification node: FCA Consumer Duty, SOX, Basel III, EU AI Act, SEC Reg D, GDPR, plus an internal framework. The same detection runs on your memos and on our own shipping policies.
  • Anonymized aggregation is opt-in. Your data never contributes to the public Bias Genome or cross-org causal weights unless the org admin flips the switch inside Settings → Privacy.

Full data-handling and sub-processor list: /privacy.

Each node’s audit trail maps onto EU AI Act Art. 14, Basel III ICAAP, and SEC AI disclosure.See the mapping
Your turn

Run the engine on your next strategic memo.

Upload takes 60 seconds. The twelve-node pipeline you just read about runs on your document and returns a DQI, flagged biases, toxic combinations, and the questions the memo didn't ask.

Audit your memo See the proof Read the R²F standard