# Agencies Reassess Legacy Software for AI

Google Cloud executive Jeanette Manfra argues that legacy enterprise software, vendor lock-in, and ongoing patching burdens can slow government AI adoption and increase cybersecurity risk. A cited survey found that 78% of respondents say legacy maintenance diverts resources from AI innovation, while 75% say existing enterprise software limits their ability to use AI. For procurement professionals, these findings highlight how software choices and licensing terms can affect agencies’ ability to modernize securely and adopt AI.

- Agencies and contractors should assess interoperability and licensing flexibility when planning software modernization and acquisitions, as these factors can influence vendor lock-in and future AI adoption.
- Acquisition planning can account for secure-by-design requirements and AI-enabled vulnerability remediation to address the security risks associated with patching burdens.
- The survey results indicate that maintaining legacy systems competes with resources for AI innovation, making modernization tradeoffs relevant to agency technology and cybersecurity planning.

**Jurisdictions:** federal
**Industries:** Information Technology
**Topics:** Artificial Intelligence, Cybersecurity
**Published:** November 19, 2026

### Government Entities
- Cybersecurity and Infrastructure Security Agency (CISA)
- Department of Homeland Security (DHS)

### Vendors
- Google Cloud (AI and cloud technology provider; source of cited research)
- Microsoft (Enterprise software provider discussed in relation to legacy software and security patching)

### Sources
- [
        When the legacy software meant to delivery AI becomes a trap - Federal News Network](https://federalnewsnetwork.com/commentary/2026/11/when-the-legacy-software-meant-to-delivery-ai-becomes-a-trap) - Federal News Network