Alation on Monday unveiled AI Governance, a new suite purpose-built to help enterprises remain regulatory compliant as they move agents and other AI tools into production.
After experimenting with AI over the past few years, many organizations have refined AI applications to the point that they can be trusted to deliver accurate outputs. However, as new AI regulations emerge and others change, compliance is a challenge.
Alation, a metadata management specialist that already provides certain AI governance capabilities, is now adding AI Governance to directly address compliance by providing customers with a system of record to demonstrate their adherence to government oversight.
Components of the suite include a registry to inventory an organization's AI models and applications, cards that display vital information about each model -- including applicable regulatory requirements with all fields citing their sources -- and a dashboard for executives to monitor compliance.
Given that AI Governance addresses a growing problem for many enterprises, it is a valuable addition for Alation users, according to Stewart Bond, an analyst at IDC.
"It is significant because it addresses a genuine and growing pain point -- the manual, fragmented process AI and data leaders in organizations face when trying to demonstrate AI compliance to boards and regulators," he said.
With the responsibility for an organization's AI governance increasingly being assigned to data leaders, which is the target audience for Alation's data catalog and other data intelligence capabilities, the new suite is a logical addition for the vendor, Bond continued.
"We have seen a shift of AI governance responsibility toward organizational data leaders, which also plays directly to Alation's existing strengths in data intelligence, making this a natural and well-timed extension of its core platform," he said.
Keeping compliant
With successful AI development comes new problems.
It is significant because it addresses a genuine and growing pain point -- the manual, fragmented process AI and data leaders in organizations face when trying to demonstrate AI compliance to boards and regulators.
Stewart BondAnalyst, IDC
OpenAI's November 2022 launch of ChatGPT sparked surging enterprise interest in AI development. In response, data management and analytics vendors whose platforms oversee the data that feeds AI tools the intelligence they need to perform have built environments to simplify AI development.
Despite the investments many organizations have made in AI initiatives and the development tools provided by vendors, most AI projects continue to fail before reaching production. The reasons vary, but poor data quality and disorganized data estates that make it difficult to discover and retrieve relevant data are among them.
Now, with many enterprises placing greater emphasis on data quality and better organizing their data, and with many vendors adding tools that improve data retrieval, some are successfully deploying agents and other AI tools.
Suddenly, the new challenge for them is staying on top of evolving AI regulations and not running afoul of any of them.
GT Volpe, Alation's head of product management, noted that as Alation customers began to scale their AI initiatives, they had pieces of AI governance in place, but they were inconsistent and applied piecemeal through emails and spreadsheets. The new suite is designed to provide a consistent compliance layer.
"Alation already governs the data AI depends on -- quality, lineage, and policies," he said. "Extending that foundation to govern AI itself was a natural move. The substrate was already there. What was missing was the AI-specific layer."
Alation's compliance-specific AI Governance suite includes the following features:
AI Asset Registry to provide a comprehensive, searchable inventory of an organization's models, agents and other AI tools.
Model Cards generated from the metadata of AI assets that demonstrate -- citing sources -- whether there is documented proof that the AI tools meet regulatory requirements or whether more human verification is needed.
Agent-powered governance workflows that route high-risk AI assets to appropriate departments for remediation.
A registry of regulations that provides built-in support for statutes such as the EU AI Act, GDPR, NIST AI RMF, and ISO 42001 to guide teams as they address compliance.
A dashboard for chief data officers, chief information officers, chief revenue officers and chief compliance officers that displays the state of their organization's compliance.
Governance -- or lack thereof -- is often a hindrance as enterprises scale AI initiatives, according to Michael Ni, an analyst at Constellation Research. As a result, Alation's AI Governance is significant for the vendor's users.
"AI Governance is quickly becoming an operational bottleneck, not a documentation exercise," he said. "Enterprises are racing to deploy AI agents, but almost none can answer a regulator's simplest question: 'Show me every AI system you have, what data it uses, who approved it, and whether it complies.' Alation sees that governance gap becoming the next enterprise AI crisis."
Competitive standing
Like Alation, data intelligence providers such as Collibra and Informatica are among the many vendors now offering AI governance capabilities. In that respect, Alation's AI Governance is not differentiated, according to Bond. However, its agent-powered workflows could distinguish it to some degree.
"Alation is not first to market, and the announcement carries some catch-up character," he said. "That said, the agentic workflow routing driven by regulation applicability is a meaningful differentiator, moving beyond static cataloging toward dynamic, automated governance."
Bond added that Alation's five-component architecture for ensuring compliance with AI regulations, covering the full lifecycle from asset registration through workflow routing to executive oversight, is coherent. However, a risk-scoring engine that monitors model drift would strengthen AI Governance, he continued.
"Alation already provides an Open Data Quality Framework and references model monitoring more broadly across its platform, so the gap may be more about surfacing those capabilities explicitly within the new AI Governance offering than building them from scratch," Bond said.
Ni similarly noted that other vendors address compliance with their AI governance capabilities. However, being first to market is not what's important from a competitive standpoint, he continued. Instead, it's providing a trusted system for answering board and regulator demands.
"Alation's announced capabilities are not purely catch-up, but also not a standalone category-defining leap," Ni said. "What Alation brings and is their opportunity is connecting AI governance to metadata, lineage, business context and evidence management in a way that data teams already use to govern data."
The composition of AI Governance is a significant improvement over patchwork AI governance frameworks enterprises put in place on their own, Ni added. However, it could be improved by including capabilities that govern agent behavior beyond documentation.
"What's still missing is the runtime layer," Ni said. "Alation's announcement is strong, [but] governance also has to address … governing AI behavior while those systems are operating. As AI-driven automation and agents execute more tasks, enterprises need to shift from documentation to runtime controls, governance and accountability."
Looking ahead
Alation's product development plans are focused on delivering governance to the entire AI workflow, according to Volpe.
That includes governing data, the capabilities that contextualize data, such as semantic models, and -- as Ni suggested -- agents and other AI tools in production.
"The connective tissue between all three layers is feedback loops, human-in-the-loop checkpoints and measurement so that every decision logged and every outcome measured compounds accuracy and governance over time," Volpe said.
Logging and measuring AI outcomes would be wise, according to Bond, who suggested that Alation could improve AI Governance by adding capabilities that audit agents in action rather than stopping at how agents were approved.
"Extending the platform to cover agentic AI behavior … would address a fast-emerging governance gap that few vendors have solved," he said. "That capability would also be a credible differentiator that could attract net-new customers in sectors like financial services and healthcare, where agent accountability is becoming a hard requirement."
"A trust layer goes beyond documenting AI systems to continuously validate whether they are using trusted context, following policy, and producing explainable outcomes," he said.
Eric Avidon is a senior news writer for Informa TechTarget and a journalist with more than three decades of experience. He covers analytics and data management.
