From Data to Business Value: Unlocking the Power of Data Products

Enterprises are awash in data. How can leaders transform all that raw data into information that delivers business value?

For many, data products are emerging as the answer. In a recent webinar hosted by CDO Magazine and Alation, industry leaders shared their insights on how modern enterprises are leveraging data products to unlock business value. Rupinder Dhillon, VP, Enterprise Data at Sobey’s was joined by Katie Chesney, Vice President of Product at 84.51, and Joe Bonanno, Global Head of Data and Analytics at Citi in a conversation moderated by myself, Steve Wooledge, VP of Product Marketing at Alation, for a discussion that shed light on the evolving role of data products, organizational alignment, and best practices for success in today’s AI-driven landscape.

In this blog, we’ll recap the key takeaways from the webinar, which you can also watch below. Let’s dive in! Some quotes have been edited for clarity. 

Defining data products: A bridge to business value

I kicked off the session by framing data products as “curated, reusable data assets that have a specific business purpose.” These products bridge the gap between raw data and actionable insights, supporting both end-users and downstream AI models and systems.

For the Canadian supermarket chain Sobey’s, Dhillon emphasized the need to embed data products into operational systems, making them actionable for decision-making. “We’re looking to take data products from being an analytics tool to something that fundamentally changes our business,” she said.

Joe Bonanno from Citi offered a historical perspective, describing how the concept of data products has evolved over time. “We would build what we call data pipelines. And ultimately we would save it as what we would call a data product, maybe 5 or 10 years ago, so that other people can have a starting point with control totals and then maybe spin it off into different permutations,” he explained. 

Today, Bonanno sees data products as encompassing more advanced tools, such as “decision engines to target the right client at the right time with the right opportunity,” which are “machine learning algorithms that can be reused” and self-learn over time. He highlighted examples like intelligent document automation, where attributes are extracted from lengthy PDFs using NLP, and intelligent assistants that function like a “little bot” capable of answering highly specific business questions repeatedly. These reusable, self-learning tools represent the breadth of what Bonanno described as “the landscape of things that are reusable and consumable, and some of them are self-learning, and some of them are, you know, one-offs.” His comments highlight the vast spectrum of types of data products leaders will see, depending on an organization’s needs and use cases, particularly in the age of AI. 

The importance of business alignment and accountability

A recurring theme was the need for tighter collaboration between data and business teams. Katie Chesney explained how Kroger and 84.51 foster accountability by involving business stakeholders early in the process. She highlighted the creation of a data maturity scorecard that measures business engagement and data quality improvements.

Business involvement is critical, Chesney emphasized. The team’s introduction of the scorecard has enabled them to move from a tech-owned model to one where the business is deeply involved in shaping data strategies and defining success metrics.

Dhillon echoed this sentiment, highlighting the importance of using business-friendly language when engaging stakeholders. “You can't use the words critical data elements,” she advised. “Let's start the conversation with the business processes that they understand.” For example, for conversations with merchandisers, ask what information they need about suppliers to meet their goals. “As data people, we have a tendency to come in with super data-heavy jargon, and that can be incredibly intimidating.” 

Bonanno shared how Citi organizes its teams into cross-functional squads, blending business and tech expertise. The key is to foster a sense of shared ownership, he said. While the data teams build the tools, it’s the business that drives their usage and ensures they deliver value, he emphasized

Governance and trust: Foundations for scalable AI

Data governance was another critical topic for the panel. As AI systems become more pervasive, the importance of trust and controls cannot be overstated. As a bank, Citi is a highly regulated environment. Bonanno described how Citi ensures robust governance for its AI initiatives, sharing how the team brings in legal, risk, and compliance partners from the start to establish guardrails and controls for AI initiatives.

Chesney, too, highlighted the role of partnerships in accelerating governance. “At first we thought we had to build everything internally, but we’ve taken a step back from that,” she shared, revealing how her team has learned to focus on what gives them a competitive edge and where to partner with best-in-class solutions. “It's not a secret. We partner with Alation for data discoverability within Kroger and 84.51,” she shared. “And it's been an amazing relationship that's helped us accelerate from a discoverability standpoint.”

Dhillon added that AI data products can act as a catalyst for improving data quality. “Data products have actually allowed us or given us an opportunity to reinforce why governance and quality are important because it's directly related to the efficacy of the algorithm and the output and the accuracy and the ability to scale it.”

Lessons learned: Building a culture of experimentation

All panelists stressed the importance of agility and adaptability when implementing data products. Chesney advocated for a “test and learn” mindset, emphasizing the value of course correction. “Where we started and where we are today are vastly different. Don’t be afraid to pivot if something isn’t working,” she advised.

Dhillon warned against the risk of data teams becoming disconnected from business needs. “It’s easy to fall into the trap of building ‘data vanity projects.’ Always tie your initiatives back to immediate business value,” she cautioned.

Bonanno highlighted the need for due diligence and early alignment. He advises that before you build, socialize the idea with stakeholders, test prototypes, and ensure alignment with legal and compliance teams. It’s better to address potential issues early than to redo work later.

Key takeaways for data leaders

The panelists agreed that the future of data products lies in their ability to power AI-driven innovation. This includes leveraging both structured and unstructured data and enriching data with semantic context for greater utility.

The webinar also offered several actionable insights for data leaders:

• Involve the business early: Use clear, business-friendly language to foster alignment and accountability.

• Invest in governance: Establish strong controls and partnerships to build trust in data products.

• Adopt a test-and-learn approach: Be willing to experiment and pivot as needed to meet business objectives.

• Leverage AI thoughtfully: Combine structured and unstructured data with semantic context to power scalable AI solutions.

By following these best practices, organizations can unlock the full potential of data products, driving both innovation and measurable business value.

Curious to learn how a data catalog can help you deliver trusted data products and AI? Book a demo today.