The Department of Veterans Affairs (VA) is exploring how data, AI, and automation could help it anticipate veterans’ needs and connect them with benefits before they have to navigate the agency’s systems themselves, according to a MeriTalk report.

VA Chief Data Officer (CDO) Lisa Rosenmerkel discussed the approach at the GovCIO Federal Cloud & Data Forum in Washington, D.C. on October 8. She said the department wants to make its data reliable and accessible enough to support more proactive services for veterans.

The approach would involve better connecting data across VA systems, improving interoperability, and exploring automation where data cannot be centralized. 

Rosenmerkel emphasized that data quality remains essential to realizing their value.

Data Quality Becomes an AI Requirement

For CDOs, as Rosenmerkel pointed out, data reliability is at the center of any AI-enabled service delivery. An AI system that identifies a potentially relevant benefit still depends on accurate, current, and connected information.

That makes data quality, lineage, access controls, and consistent definitions important requirements. 

Rosenmerkel said AI does not inherently improve data. Instead, it can make an organization more aware of its data’s limitations. VA is therefore working to ensure that veteran data is formatted in ways that can be accessed across multiple programs, reducing repetitive questions and application friction.

The VA’s approach also illustrates why data-management processes increasingly matter as agencies move from using AI for information retrieval toward systems that can trigger or recommend actions.

The challenge is to ensure that the data used to anticipate a need is trustworthy enough to support an interaction with a veteran, and that the agency can appropriately explain and safeguard any recommendation the system provides.