Dirty Data Expert
A contrarian view on AI, ERP and data integrity

50 Years of IT Promises. Why AI Risks Failing Again.

50 years of IT promised better decisions. Most delivered better systems. Dirty Data explains why.

  • 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

The central argument

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.”

Executive Brief

Before You Replace Your ERP for AI — Read This

Most organisations believe AI requires a new system.

It doesn’t. It requires better architecture.

This short executive brief explains why leading companies:

  • Keep ERP stable
  • Run AI outside it
  • Move faster with less risk
Executive Insights

Sharp thinking. No noise. Built for decision-makers.

A concise authority layer for executives who want the argument fast before they decide where to go next.

Built for Sales. Not for Decisions.

A history of smoke, mirrors, and missed decisions. Read the executive framing behind the core argument.

Read Executive Insight

Built for transactions. Not for decisions.

Many enterprise systems are exceptional at recording activity, but far weaker at guiding better choices.

Clean data is not decision-ready data.

Accuracy alone does not make data usable. Structure, segmentation and context matter just as much.

Intelligence layered onto weak structure increases risk, not clarity.

If the foundation is weak, more intelligence can amplify the weakness rather than resolve it.

About Gary Segal

Gary Segal is a data strategist, author, and advisor with a career spanning the full evolution of business systems — from manual accounting ledgers to modern AI-driven platforms. His work sits at the intersection of finance, data architecture, and executive decision-making.

A lifetime across data evolution

Gary’s experience covers the complete journey of business information systems: early accounting structures, the rise of spreadsheets, the emergence of business intelligence, OLAP, and data warehousing, through to today’s embedded AI solutions. This long-view perspective gives him a unique ability to identify patterns that others miss — particularly where systems promise decision support but fail to deliver it.

He has worked extensively in retail, FMCG, and inventory-driven environments where complexity is high, margins are tight, and decisions matter. Across these environments, one issue consistently emerges: data may be processed correctly, but it is rarely structured correctly for decision-making.

The Dirty Data perspective

Through his work and writing, Gary challenges one of the most widely accepted assumptions in modern business: that clean data is sufficient for good decisions. His central argument is clear — without proper segmentation and structural integrity, even clean data can lead to poor outcomes.

This thinking forms the foundation of Dirty Data, where he exposes how transactional systems, reporting layers, and now embedded AI solutions often reinforce flawed structures rather than correct them. The result is a growing gap between what systems report and what executives actually need to know.

His work is aimed at decision-makers who are not looking for more dashboards, but for clearer thinking, better questions, and more reliable foundations for strategy.

Embedded AI

A history of systems that promised transformation — and delivered complexity

From spreadsheets to ERP, BI, data warehouses and AI, each wave of technology promised clarity and control. Most improved operations. Few improved decisions.

The argument

Technology kept advancing. Decision quality did not keep pace.

Businesses were told that each new wave of software would produce insight, control and confidence. Too often, it produced more layers, more dependency and more cost.

The warning

Embedding AI into ERP risks repeating the same pattern

AI layered into weak data structures can accelerate output without strengthening understanding.

Embedded AI Article

ERP captured transactions with precision. It never solved how decisions are made.

A system can be operationally excellent while still failing to support the strategic choices executives need to make.

See the framework Read Embedded AI
Embedded AI Article

Better dashboards on weak structures do not create insight. They create cleaner confusion.

Reporting becomes more polished, but the thinking underneath it remains fragile.

Explore Dirty Data
Embedded AI Article

AI embedded into the wrong foundation does not fail loudly. It fails convincingly.

Weak assumptions can look sophisticated when wrapped in powerful technology.

Read the sample

The problem was never technology alone.

Over decades, organisations invested in systems that improved speed, automation and reporting. Yet decision quality rarely improved at the same rate.

The missing layer was not another platform. It was integrity — the integrity of transactions, the integrity of segmentation, and the integrity of structure required for meaningful analysis.

Dirty Data

The hidden threat to transactional and segmented integrity

Dirty Data is not about untidy records. It is about structural weakness — data that cannot be trusted, segmented or used for serious decisions.

Transactional integrity

Can the business trust the underlying records?

Without consistency and control, reporting becomes noise and decisions become guesswork.

Segmented integrity

Can the data be grouped in a way that supports decisions?

Without coherent segmentation, even clean data becomes difficult to use and easy to misread.

Sample Read

Read before you decide

What the reader will get

Dirty Data is written for executives who recognise that better tools have not delivered better decisions. It shows why the issue is not just data quality — but data structure.

  • Why IT projects repeatedly overpromise and underdeliver
  • Why ERP and reporting systems can be operationally strong but analytically weak
  • Why transactional and segmented integrity both matter
  • Why embedding AI into weak structures increases cost before value

The core message is simple: before a business adds more intelligence, it should make sure the data is trustworthy, segmentable and fit for decisions.

Inside the sample

Page 17 — Major culprits of Dirty Data

For 50 years, technology promised to improve decisions. In many cases, it improved process faster than judgment.

Page 19 — Chronological order of Dirty Data Culprits & the Rise of Segmentation

A company can have clean-looking data and still lack the structural integrity required for analysis and AI.

Page 23 — The executive message

If leadership trusts output from weak foundations, the cost is not technical. It is strategic and commercial.

Take the next step

Move from insight to framework

By this point, the problem is clear. The next step is understanding how to address it.

Start with the argument

For visitors arriving from LinkedIn or thought-leadership content, this route keeps the emphasis on understanding first.

  • Embedded AI explains why systems underdelivered
  • Dirty Data explains the structural cause
  • The homepage sample gives readers enough substance to engage seriously
Read from the top
Best for first-time visitors who need the narrative before the purchase decision.

Get the Executive Playbook

For executives who recognise the problem — and want the full framework.

  • Executive-ready framing
  • Clear stance on ERP, AI and data integrity
  • Built for leaders who need decisions, not buzzwords
Buy on Gumroad
A concise framework for leaders who want clarity before investing further in AI and analytics.

Before you embed AI, inspect the foundation.

Better systems do not automatically produce better decisions. If the data lacks transactional integrity, segmented integrity and decision-ready structure, the next layer of technology may only make the problem more expensive.

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