Edgar Randall of Dun & Bradstreet argues that UK businesses cannot unlock the promised growth of AI without first establishing a robust and well-defined data strategy.
In a relatively short space of time we’ve gone from AI being something of a pipe-dream for many, to something integrated into the very fabric of our society. The benefits of AI for businesses have been discussed at length, from analysing vast datasets to generating insights and automating processes. At this point, most businesses have at least some experience using AI, with many creating their own AI strategies, products and solutions.
For CIOs specifically, AI represents an unmissable opportunity to drive innovation, efficiency and competitive advantage. Yet, beneath its many benefits lies a critical challenge that’s not spoken about enough: AI is often opaque. Its outputs, while sometimes correct, are not infallible. And when you can’t discern when AI is wrong, trusting it to be right becomes a gamble – and not one you should risk your business on.
This opacity poses a significant risk for the many organisations and CIOs, betting on AI to deliver business value. If AI’s outputs are unreliable or misleading, the consequences can be catastrophic, ranging from operational inefficiencies at one end of the scale to significant reputational damage and loss of revenue at the other.
Addressing this challenge requires a multi-faceted approach – technical, cultural and strategic. But at the heart of any successful AI strategy lies a fundamental truth: if you want to trust what comes out of AI, you need to trust what goes into AI. In other words, there is no AI strategy without a robust data strategy.
The AI-data nexus
Put simply, AI systems, whether powered by Machine Learning, large language models, or other architectures, are only ever going to be as good as the data they process. Data is the raw material that fuels AI’s decision-making, predictions and outputs. Poor-quality data – incomplete, inaccurate, biased, or outdated – leads to flawed results, no matter how sophisticated the algorithm. Conversely, high-quality, well-governed data enables AI to deliver reliable, actionable insights.
For CIOs, this means that building trust in AI has to begin with establishing trust in data. A comprehensive data strategy is the foundation for ensuring that AI systems operate on a bedrock of accuracy, consistency and transparency.
Key pillars of a data strategy for AI
A data strategy tailored for AI success rests on several critical pillars:
1. Data quality and integrity
High-quality data is non-negotiable. This means ensuring data is accurate, complete and consistent across sources. For example, if an AI system is used to predict customer behaviour, discrepancies in customer data, such as duplicate records or missing fields, can skew predictions. Implementing robust data cleansing, validation and enrichment processes is essential to maintain integrity at every stage of the data lifecycle.
2. Data governance and compliance
Trust in AI also hinges on trust in how data is managed. A strong governance framework ensures data is collected, stored and processed in compliance with regulations like GDPR, CCPA, or industry-specific standards. Governance also establishes clear policies for data access, usage and security, reducing the risk of misuse or breaches. For AI applications, governance extends to ensuring transparency in how data influences outcomes, enabling organisations to trace and explain AI decisions.
3. Data integration and accessibility
AI thrives on diverse, unified datasets. Siloed data, scattered across departments or legacy systems, limits AI’s ability to deliver holistic insights. A data strategy must prioritise integration, creating a single source of truth that AI systems can tap into. This requires modern data architectures, such as data lakes or cloud-based platforms, that enable seamless access while maintaining security and scalability.
4. Bias detection and mitigation
AI can inadvertently amplify biases present in training data, leading to unfair or inaccurate outcomes. For instance, biased hiring algorithms have made headlines for perpetuating gender or racial disparities. A data strategy must include mechanisms to identify and address biases, such as regular audits of datasets and the use of fairness-aware algorithms.
5. Data literacy and culture
A data-driven culture is critical to maximising AI’s potential. Employees at all levels need to understand the importance of data quality and governance. CIOs should champion data literacy programmes to empower teams to contribute to and trust the data ecosystem. This cultural shift ensures that AI initiatives are supported by a workforce aligned with the organisation’s data strategy.
The business imperative
For CIOs, the stakes are high. AI investments are substantial, often involving significant financial and organisational resources. Without a data strategy, these investments are at risk of underdelivering or failing outright. Organisations with mature data strategies are better positioned to scale AI initiatives, enabling smart decision-making, enhancing customer experiences and driving operational efficiencies.
A robust data strategy also unlocks competitive differentiation. In industries like finance, healthcare, or retail, where AI is reshaping customer interactions and operational models, trusted AI outputs can set organisations apart. For example, a retailer using AI to personalise customer recommendations relies on accurate, integrated data to deliver relevant suggestions, fostering loyalty and driving revenue.
AI can be a double-edged sword, but its potential cannot be realised without a foundation of trusted data. The message is clear for CIOs: an AI strategy is only as strong as the data strategy behind it. By prioritising data quality, governance, integration, bias mitigation and literacy, organisations can unlock the full power of AI while minimising risks. In a world where AI’s outputs shape decisions and drive outcomes, trust is the ultimate currency. And that trust begins with data.
