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 AI50 years of IT promised better decisions. Most delivered better systems. Dirty Data explains why.
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.
Executive Brief
Most organisations believe AI requires a new system.
It doesn’t. It requires better architecture.
This short executive brief explains why leading companies:
A concise authority layer for executives who want the argument fast before they decide where to go next.
A history of smoke, mirrors, and missed decisions. Read the executive framing behind the core argument.
Read Executive InsightMany enterprise systems are exceptional at recording activity, but far weaker at guiding better choices.
Accuracy alone does not make data usable. Structure, segmentation and context matter just as much.
If the foundation is weak, more intelligence can amplify the weakness rather than resolve it.
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.
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.
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.
From spreadsheets to ERP, BI, data warehouses and AI, each wave of technology promised clarity and control. Most improved operations. Few improved decisions.
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.
AI layered into weak data structures can accelerate output without strengthening understanding.
A system can be operationally excellent while still failing to support the strategic choices executives need to make.
See the framework Read Embedded AIReporting becomes more polished, but the thinking underneath it remains fragile.
Explore Dirty DataWeak assumptions can look sophisticated when wrapped in powerful technology.
Read the sampleOver 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 is not about untidy records. It is about structural weakness — data that cannot be trusted, segmented or used for serious decisions.
Without consistency and control, reporting becomes noise and decisions become guesswork.
Without coherent segmentation, even clean data becomes difficult to use and easy to misread.
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.
The core message is simple: before a business adds more intelligence, it should make sure the data is trustworthy, segmentable and fit for decisions.
For 50 years, technology promised to improve decisions. In many cases, it improved process faster than judgment.
A company can have clean-looking data and still lack the structural integrity required for analysis and AI.
If leadership trusts output from weak foundations, the cost is not technical. It is strategic and commercial.
By this point, the problem is clear. The next step is understanding how to address it.
For visitors arriving from LinkedIn or thought-leadership content, this route keeps the emphasis on understanding first.
For executives who recognise the problem — and want the full framework.
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.