How AI Agents Are Revolutionizing Data Catalogs and Governance

By Talo Szem

Published on March 28, 2025

Enterprises are navigating a profound data management crisis. Forrester Research found that "less than 0.5% of all data is ever analyzed and used" and projected that if the average enterprise could boost data accessibility by just 10%, it would generate more than $65 million in additional net income.

Data leaders, in turn, are under increasing pressure to navigate this crisis. Modern leaders must demonstrate how data management can deliver business outcomes. This raises a critical question: how can software evolve to support them?

At a recent webinar hosted by Alation, industry experts Satyen Sangani, CEO and founder of Alation; Susan Wilson, Sales Leader at Alation and former data leader at Pfizer for 15 years; and Greg Sands, managing partner at Costanoa Ventures, offered an answer. The trio discussed the urgent need for a transformation in software and data management practices – and shared insights on how Alation is evolving to both incorporate and support the creation of AI agents.

From traditional catalogs to AI-driven data management

Metadata has long been heralded as a key ingredient to optimizing data management. By sharing context on data usage, metadata builds trust and confidence while reducing duplicative work. Traditional data catalogs emerged primarily as IT-driven tools for managing metadata. However, manual processes often led to limited visibility and insufficient use.

Sangani recounted his experience at Oracle, which inspired him to launch the first modern data catalog. At the time, he ran a line of business that sold financial analysis apps to finserve companies. It was a painful, tedious process. “Often, implementing these applications took literally years just to acquire data and format it for use,” he recounts.

Greg Sands echoes his sentiment, pointing out that despite decades of attempts, traditional metadata management projects often resulted in costly failures. "A big bank spent $100 million on a metadata management project—$3 million on software, $97 million on services—and generated reports that were never used again."

Metadata presented a solution. “It struck me that these were fundamentally knowledge management problems around data,” says Sangani. “Everybody wanted data, yet nobody knew how to use it—it was never clean, never high-quality enough, never accessible enough."

Inspired by successful information-sharing platforms like Wikipedia and LinkedIn, Sangani envisioned Alation as a social data platform, democratizing accessibility and making metadata visible and usable beyond just IT. The founding team realized, “We really can solve these problems by leveraging some of the very primitive artifacts of the internet, which is really about information dissemination,” he shares. Ultimately, this idea of social data sharing evolved into the data catalog category.

Yet the landscape has shifted once more. Today's technology ecosystems have grown in volume, adding complexity to processes. An intervention is essential. The answer? AI agents. They are poised to play a pivotal role in automating and streamlining these processes.

Tackling the modern data management crisis

Today's data teams face unprecedented pressure. The technology sector has become fiercely competitive and budgets more tightly scrutinized. As a result, leaders must increasingly justify their ROI. “For the last 10 to 12 years, with the advent of Hadoop and, of course, self-service BI tools like Tableau, there was the secular investment in data where you would just always get budget, the budget would always be bigger, and people would see the CDO as this inevitable path to the C-suite,” Sangani points out.

Today’s environment is remarkably different. Now, with high interest rates and intense scrutiny of subscription costs, data teams face a perfect storm of pressure. The question has become: 'How do we deliver value faster and better?'

The rise of AI has only made this challenge more pressing. "Leaders feel both fired up and freaked out,” says Wilson, who spent nearly 15 years at Pfizer leading data initiatives. While AI can supercharge productivity, concerns around data accessibility and quality linger, with about half of all data leaders citing these obstacles as their two biggest blockers in conversations.

Agentic AI can help leaders overcome these obstacles. Wilson highlights how enterprises are actively responding to these challenges using Alation. “Teams are tearing up old playbooks and reinventing themselves,” she says. For instance, a large U.S. airline is leveraging Alation to create end-to-end customer experiences. A major oil and gas company is using Alation and Snowflake to drive operational efficienciesand an American insurer is building stewardship agents for critical data elements.

While it’s still early days, these stories reveal how agentic AI can ease the pain of data management, and offer relief to overwhelmed leaders. By leveraging AI, such leaders can navigate this crisis, using automated solutions to directly drive business outcomes.

How AI agents transform governance and compliance

AI agents are specialized AI programs designed to both know things and do them. In the world of data management, they can automate tasks traditionally performed by humans, such as compliance and metadata management. AI agents can be tailored for structured, repetitive processes, offering precise and efficient handling of complex data tasks.

The landscape has made AI agents essential. Sands points out that data volumes have increased 100 times since Alation was launched 13 years ago. Regulatory complexity has grown five times more onerous. And budgets, which once grew at 20% a year, are now shrinking by 10% annually. For all these reasons, “We can no more think of ourselves as enabling platforms,” he says. “We need to be in the business of delivering value and business outcomes.”

AI agents can realize this vision. Sangani points out that AI agents can gather and use contextual knowledge and best practices, turning data management from procedural to declarative.

Context such as, “What is the company's current position and level of data maturity? What are the best practices and where do those things live and who knows them?” can be fed to the models. “You can take all of that context and turn data management from this procedural thing where I have to know MDM and data cataloging and metadata management and all of these things… to something that is really just declarative: ‘Just let me comply with GDPR. Just get me the data that allows me to build a better churn model,” Sangani illustrates.

In this way, AI agents present a compelling opportunity for organizations to leverage metadata context, so they may streamline compliance and governance efforts, freeing resources to focus on strategic initiatives. “People don't want to do governance manually—they just want outcomes,” Sangani concludes.

Launch of the Alation Agentic Platform

In response to these evolving demands, Alation has introduced the Agentic Platform, designed to make data catalog functionality invisible yet deeply integrated into everyday workflows.

"I've spent the last decade building the catalog; I'll spend the next few years making it invisible,” Sangani says. AI agents built directly into the platform will support this work. And on the flip side, new catalog features empower developers to build their own enterprise AI agents, using internal data, as well.

Alation’s Agentic Platform includes the pioneering MCP-based AI Agent SDK. This kit enables organizations to build customized, AI-powered solutions that leverage trusted, governed data. With this SDK, data teams can rapidly innovate and solve complex data management challenges without extensive coding or technical intervention.

The webinar also introduced Alation's Data Product Marketplace and its new Data Quality solution, designed explicitly to support AI and agent-driven processes. These products streamline discovery, automate data quality, and ensure that AI models consume reliable, governed data.

Adoption is quickly on the rise. "Organizations are using data products and AI agents to achieve tangible outcomes—enhancing customer experiences, improving efficiency, and simplifying regulatory compliance,” says Wilson.

A vision for the future: Invisible data management

The webinar underscored a powerful vision for the future—where AI-driven agents automate data catalog and governance tasks, making data accessible while freeing human talent for strategic, value-driven activities.

Sands summarizes the transformative potential, pointing to data products as a key enabler: "Data products must now be built to be consumed by machines, not just humans. That means higher quality, trustworthiness, and minimal human intervention."

As the role of data catalogs continues evolving, embracing AI and agents will be key for enterprises aiming to stay competitive, compliant, and innovative.

To learn more: