NewsletterDIC™ issues CasesAutopsies · Board WorkPractice · Research RoadmapBoard · CEO EngagementsAdvisory work Subscribe free →

// Decision Autopsies · Boardroom Reality Checks

Real failures. Every layer mapped.

Six decision autopsies. Five boardroom reality checks. Each one traces a real institutional failure through the DIC™ — showing exactly which layer broke, when, and why nobody saw it coming.

// Decision Autopsies — The Record Shows

DA-2025-001 · Credit · UK GSIB
640 loans approved. Every dashboard green. Credit committee signed off. The exposure was invisible — until it wasn't.
The model scored correctly. The rules were followed. The intent was never encoded. L3 was absent for 14 months before anyone asked the right question.
DIC™ layer that broke: L3 Intent — L4 Rules (drift between intended and executed policy)
Read the full autopsy on Substack →
DA-2025-002 · Home Health · US Healthcare
Scheduling algorithms optimised for efficiency. Patient outcome indicators deteriorated across an entire region. No single decision was wrong.
Every individual scheduling decision was defensible. The emergent pattern — resource concentration, outcome drift — was visible only at L7. Nobody had designed a feedback loop.
DIC™ layer that broke: L7 Outcome — L8 Feedback (no mechanism to detect emergent harm)
Read the full autopsy on Substack →
DA-2025-003 · Acute Care · Hospital Network
Triage protocols updated. Clinical judgment delegated to a scoring engine. Outcomes shifted — not dramatically, but consistently and in one direction.
The scoring engine was validated. The delegation was not auditable. L5 Judgment was fully automated without an authority boundary being defined.
DIC™ layer that broke: L5 Judgment (authority scope undefined at delegation)
Read the full autopsy on Substack →
DA-2025-004 · Onboarding · Asian Digital Bank
KYC pass rates were improving. Fraud rates were also improving. Regulatory findings arrived two quarters later about a segment that had been systematically under-screened.
The rules were consistent. The intent behind the rules had changed — but only in board documents. L3 was updated in policy. L4 was not re-encoded. The gap was 8 months wide.
DIC™ layer that broke: L3–L4 gap (intent updated in policy, not in rules)
Read the full autopsy on Substack →
DA-2025-005 · Pricing · Regional Insurance Group
Algorithmic pricing delivered margin improvement. Three segments were systematically disadvantaged. The pricing model had never been reviewed for fairness.
L1 Purpose stated "fair outcomes for customers." L4 Rules optimised for margin. The contradiction had been operating for 22 months before a regulatory inquiry surfaced it.
DIC™ layer that broke: L1–L4 contradiction (purpose and rules in direct conflict)
Read the full autopsy on Substack →
DA-2025-006 · Agentic Credit · Cross-border Platform · In progress
Seven agents. Three jurisdictions. Seven seconds. One outcome that no institution authorised.
Based on the scenario constructed in Paper 3 — When the Protocol Decides. Publishing shortly.
DIC™ layers that broke: L3 · L4 · L5 · L8 — simultaneously, across institutions
Subscribe to be notified →

// Section above

Decision Autopsies — composite cases constructed from documented institutional failure patterns. Not tied to any specific named institution beyond what is explicitly stated. Analytical and educational in nature.

// Section below

Boardroom Reality Checks — practitioner intelligence published on Substack. Questions drawn from pattern recognition across 25 years of institutional advisory work. Subscribe to read the full analysis.

// Boardroom Reality Check — Five published cases

What boards actually ask when something can no longer be explained.

Note #1 · Credit · UK GSIB
"Which loan approved in the last 12 months could your team not fully explain today — and why not?"
£48m facility. Decision reconstruction. The answer revealed a governance chain operating without an owner for 14 months.
Read on Substack →
Note #2 · Deposits · Asia-Pacific Bank
"Which deposit segment are you treating as stable today that may not be — and what changed your confidence?"
SME operating balances behaving like tactical liquidity. No threshold breached. The funding model had already changed.
Read on Substack →
Note #3 · AI Decisioning · Banking
"Name one AI-assisted decision made this month that your board could defend end-to-end if a regulator asked tomorrow."
Most boards cannot. Not because the AI failed — because the governance chain was never designed for AI-speed decisions.
Read on Substack →
Note #4 · Risk Model · Insurance
"When did your institution last ask whether its risk model still governs the institution it has become?"
The model was built for an institution that no longer exists. The gap had been compounding for three years.
Read on Substack →
Note #5 · Conduct · Retail Banking
"Which customer outcome in the last quarter was nobody's explicit decision — and who owns it?"
Emergent outcomes from compliant individual decisions. No one chose the result. Everyone followed the process.
Read on Substack →
Note #6 is in progress. Subscribe to receive it when published. Subscribe on Substack →

// Disclaimer & Legal Notices

The views, analyses, and perspectives expressed on this site are solely those of Deepak Aggarwal, presented in a personal and independent capacity. They do not represent or reflect the views, policies, or positions of any current or past employer, client, organisation, or affiliated entity.

All institutional case references — including but not limited to SVB, Knight Capital Group, Wirecard AG, Credit Suisse, Wells Fargo, Coutts, Orpea Group, Kaiser Permanente, and NHS entities — are cited solely on the basis of publicly documented regulatory findings, official investigations, court records, parliamentary reports, and other published primary sources. No non-public information has been used. All analysis is independent, educational, and analytical in nature.

Nothing on this site constitutes legal, regulatory, financial, investment, or professional advice of any kind. The Decision Integrity Chain™, Decision Engineering™, FUSE™, STAGE™, and related frameworks are proprietary intellectual property of Deepak Aggarwal. Unauthorised reproduction or commercial use is prohibited.

Case studies and scenarios described as "constructed" or "composite" are hypothetical illustrations based on documented failure patterns. They do not refer to any specific institution, transaction, or individual beyond what is explicitly stated.

© Deepak Aggarwal 2025–2026. All rights reserved. Decision Engineering™ · Decision Integrity Chain™ · DIC™ · FUSE™ · STAGE™ are trademarks of Deepak Aggarwal.