Dirty Data Expert
Gary Segal · Executive Intelligence on AI, ERP and Decision Integrity

Structure before intelligence.

Why AI Fails Before the First Model Is Built

Most AI projects do not fail because the model is weak. They fail because the business assumptions, data structures, segmentation, governance and ERP foundations beneath the model were never tested.

Join 1,000+ executives getting one 5-minute brief every week — where your data will break AI before AI breaks you.

  • ERP systems excel at transactions — not at decision-making
  • Embedding AI into weak data structures increases cost more easily than value
  • Without integrity, intelligence becomes amplification — not insight
Get the Executive Briefs Run AI Risk Diagnostic
Gary Segal

Gary Segal

Author of “Dirty Data” and “Switching ERP for Embedded AI”. Advisor and speaker on AI readiness.

The central argument

Executive Briefing Platform

For decades, businesses invested in better systems and expected better decisions. What improved was processing — not judgment.

Embedded AI explains the pattern. Dirty Data exposes the structural cause.

“We do not have an AI problem first. We have a data integrity problem first.”
Dirty Data book cover Switching ERP for Embedded AI book cover Author of “Dirty Data” • Author of “Switching ERP for Embedded AI” • Keynote speaker on AI readiness • Read by 1,000+ subscribers weekly
Executive Briefs by Email

Short, practical briefings on dirty data, governance and AI readiness, straight to your inbox. Sign up and get the Why AI Fails brief instantly.

Get the Executive Briefs
The Hidden Assumption

Most organisations assume their ERP contains an understanding of their business.

In reality, the ERP usually contains a record of transactions. Understanding the business requires categories, hierarchies, relationships, governance and segmentation.

Gartner predicts that through 2026, organisations will abandon 60% of AI projects unsupported by AI-ready data. (Gartner, Feb 2025)

What the ERP records

  • Products
  • Customers
  • Suppliers
  • Transactions
  • Inventory movements

What AI needs to understand

  • Relationships
  • Substitution effects
  • Demand behaviour
  • Business definitions
  • Decision ownership

The ERP Assumption

The business does not exist inside the transaction alone. The business exists in the relationships between transactions.

See the Hidden Assumptions
Executive Learning Path

Understand why AI fails before investing further in AI.

Follow the executive path from hidden assumptions to structure, dirty data, segmentation, governance and AI risk assessment.

Why AI Fails

Most AI projects fail before the first model is built.

See Why AI Fails →

Executive Assumptions

The hidden assumptions executives make before AI fails.

Uncover the Assumptions →

The Structure Gap

AI cannot understand a business that has never been structured.

Explore the Structure Gap →

Dirty Data

Clean-looking data is not the same as decision-ready data.

What Dirty Data Really Is →

Segmentation Imperative

Segmentation is the operating system beneath AI.

Why Segmentation Matters →

Inventory Intelligence

Inventory intelligence begins with segmentation, not technology.

Read on Inventory Intelligence →

Human Governance

AI can recommend. Humans remain accountable.

Who Stays Accountable →

Data Governance

Governance becomes operational when accountability is clear.

Make Governance Operational →

Embedded AI

Embedded AI can create executive illusion before executive intelligence.

Spot the Executive Illusion →

AI Risk Assessment

Test your data before AI tests your business.

Assess Your AI Risk →

First Principles of AI Readiness

Sharp thinking for executives navigating AI and dirty data risk.

Read the First Principles →

Executive Briefs

Download executive briefings on AI, governance and decision integrity.

Browse the Brief Library →
AI Risk Assessment

Test your data before AI tests your business.

Most AI failures begin inside business structures long before the first model is deployed. Run the diagnostic to identify dirty data, weak segmentation, governance gaps and structural risk.

What the diagnostic checks

  • Dirty data exposure
  • Duplicate records
  • Missing values
  • Weak segmentation
  • Structural inconsistency

What you receive

  • AI risk report
  • Executive summary
  • Segmentation risk indicators
  • Dirty Data Exposure Score
  • Recommended next actions
Get the AI Risk Diagnostic How the Diagnostic Works
Speaking

Book Gary Segal to Speak

Signature talk: “Structure before Intelligence” — why AI fails before the first model is built, and what leadership must fix first. For leadership teams and conferences.

Why AI Fails Before Deployment

The structural causes of AI failure: dirty data, hidden assumptions and the structure gap.

Dirty Data & Decision Integrity

Why clean-looking data is not decision-ready data, and what that costs the business.

Governance Before AI

Segmentation, ownership and human accountability as the foundations beneath AI.

Enquire About Speaking
Books by Gary Segal

The thinking behind the briefs.

Cover of Dirty Data by Gary Segal

Dirty Data

The Hidden Threat to Transactional and Segmented Integrity

Available on Amazon
Cover of Switching ERP for Embedded AI by Gary Segal

Switching ERP for Embedded AI

The Hidden Cost of Changing the Backbone of Your Business

Available on Amazon
Executive Brief Library

Download executive briefings on AI, dirty data and decision integrity.

Use these briefings to frame executive conversations before investing further in AI, analytics, automation or governance initiatives.

Why AI Fails

Why weak structure causes AI failure before the first model is built.

Segmentation Imperative

Why segmentation is the foundation beneath forecasting, inventory intelligence and AI.

Executive Assumptions

The hidden assumption executives make before AI, analytics and automation failure

View Executive Brief Library
Concerned about stock risk?
Message Gary directly. Ask what your inventory data may be hiding.