An unsung hero of AI: why data foundation matters

AI is rapidly becoming a competitive differentiator across industries. Across fast moving industries such as Consumer Packaged Goods (CPG), from assortment, forecasting and product inventory optimization to personalized promotions and pricing, AI promises faster answers.

Yet many executive teams find that despite significant investment, AI initiatives struggle to deliver consistent value, scale or trust. While organizations rush to launch AI initiatives, Gartner predicts that through 2026, organizations will abandon 60% of AI projects unsupported by AI-ready data.

The issue is not the AI itself. More often, it is the data foundation that AI is relying on. Nearly 63% of organizations are unsure if they have right data management practices for AI. In the industries where margins are tight, volumes are high, and decisions must be made at speed, AI is only as effective as the data it relies on. Without clean, harmonized, and consolidated data, AI becomes a risk rather than an advantage.

For C-suite leaders, this raises a fundamental concern: how to trust AI-driven decisions, how to prepare their organization for an AI era?

There are three well-known AI behaviours that help to explain why data foundations matter so much.

1.Sycophancy

AI systems are designed to recognize patterns and respond based on the information they receive. AI models learn patterns from data, not designed to challenge the validity of that data. When underlying data reflects incorrect assumptions or inconsistent business definitions, AI tends to reinforce those perspectives. If a product attribute is mis-segmented, or if metrics are inconsistently labelled, AI will not detect or correct these issues. Instead, it will reinforce them across insights.

2. Anchoring

AI systems are quick to establish foundational assumptions based on their initial data inputs. When this data is incomplete or inaccurate, these early insights become reference points, or ‘anchors’, that shape the future predictions and recommendations. If there is significant variation in quality or consistency across different systems, AI is likely to pick up flawed historical information and embed those inaccuracies throughout its ongoing analyses.

3. Garbage In, Garbage Out (GIGO)

The principle of garbage in, garbage out may be decades old, but it has never been more relevant. When data contains errors, duplicates, or conflicting records, AI scales those issues at speed.

These behaviours highlight a simple truth: AI amplifies whatever foundation it is built on. When data is clean and consistent, AI generates insights and recommendations that can be trusted. When the underlying data is flawed, AI can inadvertently accelerate risk rather than delivering the promised benefits.

Today, most organizations have data flowing in from every corner, retailers, customers, digital platforms, partners, and far-reaching markets. It is only when this data is cleansed, harmonised, and integrated that AI can truly amplify value. The difference between merely collecting information and harnessing it for competitive advantage lies in the reliability of the underlying data foundation.  

Around 43 % of organizations cite data quality as the main reason their AI projects stall, AI models cannot deliver reliable outputs without clean, well-structured inputs. Even small errors in data can lead to significant consequences such as having too much stock, sending incorrect pricing signals, or running ineffective promotion campaigns.  

One principle remains central in our industry: we can’t compromise on quality.  

Quality data is the foundation of trust, giving us confidence in the results we get. When data is clean, AI tools are better equipped to do a much better job by providing more accurate performance summaries and forecasts. There are fewer mistakes from AI-generated answers, and analysts can trust that what they learn is based on reliable facts. 

One of the most common barriers to AI success is misalignment across functions. AI does not naturally bring teams into alignment. This means that if different parts of a business use different definitions, AI-generated answers will simply reflect those differences.  

For AI to be truly useful and trusted across the organisation, it is important that everyone is working with the same definition and harmonised data. Carlsberg showed that creating a single, unified data foundation, where product attributes, metrics and time periods are standardized, helps the whole business fuel growth. When various strategic business areas such as commercial strategy, revenue growth management, innovation, and finance can rely on one trusted source of truth, the alignment helps teams work better together and make informed strategic decisions. 

Across industries, AI is quickly moving from experimentation to execution. As organizations look to scale AI across planning, operations and commercial campaigns, one factor consistently determines success: the reliability of the data foundation. 

An AI-ready data foundation has become a strategic investment rather than a technical one. Without this foundation, AI risks reinforcing inconsistencies and driving decisions based on fragmented or biased data. 

For C-suite leaders, strong data foundations support more accurate forecasting and planning, tighter alignment between sales, operations, and finance, and more confident decision-making. They also make it possible to scale AI across brands, channels, and regions without increasing complexity. 

Most importantly, trusted data accelerates value. It allows organizations to adopt new AI capabilities and move beyond pilots to impact at scale. 

AI technology will continue to evolve. But competitive advantage will not come from technology alone. It will come from organizations that invested early in getting their data right. 

For leaders, the question is no longer “Should we invest in AI?”
It is “Which components of our data foundation must evolve to support AI-ready data with confidence?  

In an industry defined by speed and scale, AI success starts with data you can trust. 

Discover how our data harmonization services can help you build a solid data foundation to prepare your organization to be future ready. Or contact us to learn how we simplify complexity for your business.