AI Moves to the Top of the State Technology Agenda: What It Means for HFAs

For the first time, artificial intelligence ranks as the top priority of state chief information officers, displacing cybersecurity after 12years in that position. The 2026NASCIO State CIO Survey shows that state governments have moved well beyond experimentation, and housing finance agencies are beginning to follow with applications of their own.

AI Becomes State Government's Leading Technology Priority

According to NASCIO, 82 percent of state CIOs now report generative AI projects in production, and 35 percent report enterprise-scale projects across the executive branch. Governance is expanding alongside adoption: 78 percent have adopted an AI governance framework, up from 65percent in 2025. Agentic AI, which completes multistep tasks rather than only generating content, is already in use in 29 percent of states and planned in another 35 percent. CIOs identify workflow and process automation as its most common application.

The survey also identifies the constraint that will shape adoption at every agency. Only 57 percent of states have implemented data governance for generative AI that covers data sources, quality, bias, and privacy.

How Idaho Applied AI to Borrower Service

On October 5, NCSHA recognized Idaho Housing and Finance Association with its Management Innovation Technology award for Jenna, a bilingual AI voice agent. According to IHFA's award entry, the agency receives about 4,000 inbound calls per day, rising to 7,000 or more on peak payment days. Jenna verifies each caller's identity against the loan servicing system in real time and routes the call to the appropriate team in English or Spanish.

IHFA reports that average call time fell from about 7.5minutes to about 6 minutes, and that Jenna costs about $0.06 per minute compared with about $1.00 per minute for a live agent. IHFA built the API integrations connecting Jenna to its servicing system in-house, so each verification draws on the agency's own loan records.

Governance Requirements Are Arriving with Adoption

HFAs that sell or service loans for the government-sponsored enterprises now face explicit AI governance expectations. As of March 3,Freddie Mac requires sellers and servicers to demonstrate procedures for mapping, measuring, and managing AI risk, as NCSHA reported in its March27 Washington Report. Fannie Mae's April 8 Lender Letter, summarized in NCSHA's April 17 Washington Report, requires comparable policies and extends those protections to vendors and subcontractors that use AI on a seller's or servicer's behalf.

Implications for Agency Systems

The common thread across the NASCIO findings, the Idaho example, and the enterprise requirements is that useful AI depends on structured, reliable, and well-governed data. An agency whose program, compliance, and inspection records are scattered across spreadsheets and disconnected tools will find it difficult to deploy AI responsibly, regardless of the model it selects.

ProLink's software is designed to provide that foundation. ProLinkHFA brings multiple programs and funding sources into a single architecture, Procorem centralizes documents and communication with development partners, and Procorem Inspect captures inspection results in a consistent, structured format. We will explore these questions further at ProLink Technology Live 2026, our virtual conference on November 4–5, including a session on Data Views and Dashboards. We would welcome a conversation about how your agency is preparing its data for AI.