AI Agents Raise the Stakes for Data Trust
Mike Vizard speaks with Martin Zahumensky, CEO of Ataccama, about why data trust for AI is becoming a priority for enterprise technology leaders. Zahumensky says AI agents are amplifying data issues that have existed for years. The difference is that agents can now act on incomplete, incorrect or poorly governed data at much greater speed.

That shift is forcing organizations to rethink data management. Human users often bring context, judgment and history to reports and dashboards. AI agents may not have that same business understanding. If the data behind an answer is wrong, the agent can make a flawed decision and magnify the impact across a process.
Trust Requires More Than One Data Tool
Zahumensky describes Ataccama as a trust layer for enterprise data. The company’s platform spans data cataloging, data quality, observability, reference data management and master data management. He argues that data trust for AI depends on understanding where data lives, what it means, how it is used and whether it is fit for a specific use case.

The discussion also explores why point solutions can only go so far. Many large organizations have fragmented data landscapes built over many years. AI agents may need to work across source systems, analytics platforms and operational workflows. To make those agents useful, companies need reliable metadata, lineage, quality checks and governance that can be exposed to AI systems.
Compliance and Automation Need Auditability
As AI agents become part of enterprise processes, auditability becomes more important. Zahumensky notes that organizations will need to prove which data was used, how an AI system made a decision and whether the data was appropriate for that task. That matters for regulatory compliance, internal governance and business trust.

His advice is to focus on the broader picture. Companies should not only modernize data infrastructure or chase a new AI platform. They should prepare the data foundation that makes AI useful and trustworthy. Data trust for AI starts with cleaning, documenting, governing and understanding the information that agents will use.