Artificial Intelligence (AI) is transforming industries, but its success depends on more than just powerful models. To scale AI effectively, organizations must first scale trust in their data — and that begins with building a strong data culture.
AI may dominate headlines, but behind every successful deployment is something less glamorous and far more critical: a foundation of reliable, well-managed, and trusted data. Without this, even the most advanced AI solutions will fail to deliver value.
Here are five key actions IT leaders can take to create an AI-ready data culture.
1. Treat Data as a Product, Not a By product
One of the biggest cultural shifts organizations must make is to view data not as leftover “exhaust” from business operations but as a valuable product in its own right. This means applying product management principles to data: defining ownership, standardizing formats, implementing version control, and planning for lifecycle management.
Mike Kreider, CIO of DHL Supply Chain North America, explains how his team applies this mindset. “A data product is a standardized dataset from one or more systems, formatted for easy reuse,” he says. For example, shipment data products not only support logistics and operations but also power generative AI tools like DHL’s proposal generator. If the data isn’t clean or properly managed, the tool simply doesn’t work.
Kreider emphasizes that ownership is key. Each data product has a business owner responsible for keeping it accurate and relevant, ensuring that it never becomes “orphaned.” This accountability is what keeps data trustworthy for AI applications.
IBM also follows this approach by making self-service data products available across the enterprise. As Dinesh Nirmal, SVP of IBM Software, notes: “If teams can’t easily find and trust the right dataset, they can’t innovate at speed.”
2. Build Observability and Traceability Into Data Practices
Trust in AI depends on trust in data. That means organizations need not just high-quality data, but also full visibility into where it comes from, how it’s transformed, and how it’s used. Observability and traceability provide the audit trail that makes this possible.
Dun & Bradstreet has built observability into its DNA, monitoring over 85 billion data quality points with tools like DataShield and DataWatch. DataShield enforces standards at the point of entry, while DataWatch tracks data quality over time. Together, they enable teams to identify issues, implement fixes, and confirm improvements.
Andy Crisp, SVP of Global Data Strategy at Dun & Bradstreet, stresses the importance of traceability: “If I can’t trace it, I can’t trust it.” With over 600 million business records from 30,000 sources, ensuring traceability is the only way to guarantee customers receive accurate insights.
The company also closes the loop by feeding observability findings back to regional teams and using customer feedback to validate improvements. This cycle ensures data quality isn’t just maintained, but continuously enhanced.
3. Empower Business Ownership of Data
To build a sustainable data culture, ownership must extend beyond IT teams. Business leaders need to be directly accountable for the data they generate and use. Assigning business owners ensures data products are aligned with strategic goals and remain relevant to operations.
When business leaders share accountability, data becomes a strategic asset instead of an afterthought. This shared responsibility also helps break silos and ensures data is designed for enterprise-wide use.
4. Make Trusted Data Easily Accessible
An AI-ready culture requires democratization of data. Employees at all levels should be able to easily access and trust the datasets they need without wasting time searching or validating.
Companies like IBM enable this by cataloging governed, ready-to-use data products. This reduces friction for AI engineers and analysts, allowing them to focus on building solutions rather than chasing down inputs.
5. Create Feedback Loops to Improve Continuously
Finally, data culture must be dynamic, not static. Continuous improvement is vital to ensure that data remains relevant and trustworthy as business needs evolve. Feedback loops — from internal teams as well as customers — ensure that quality improvements are meaningful and impactful.
Organizations like Dun & Bradstreet have institutionalized these loops by connecting regional teams, global strategy, and client insights. This holistic approach ensures that improvements are not just technical fixes, but changes that deliver real business value.
Building Trust for an AI-Driven Future
AI cannot thrive without a strong data culture. Treating data as a product, prioritizing observability, assigning business ownership, enabling easy access, and building continuous feedback loops are the five critical steps to ensure AI is built on a foundation of trust.
By embedding these practices, enterprises can unlock the true potential of AI — not just in theory, but in measurable business outcomes.
