The SHRM Report Is In: What AI Is Actually Doing to HR Teams in 2026

SHRM's 2026 State of AI in HR report surveyed 1,722 HR professionals in December 2025, making it one of the most comprehensive snapshots of AI adoption in people functions available today. The headline finding cuts through the hype: more than half of organisations (54%) have not implemented any AI in their HR function and have no plans to do so in 2026. At the same time, 92% of CHROs expect greater AI integration in the workforce and 87% forecast increased adoption within HR processes. The gap between executive expectation and front-line reality is the defining challenge for HR professionals this year.

Where AI is being used, adoption is concentrated rather than scattered. Recruiting leads with 27% of organisations using AI there, followed by HR technology management (21%), learning and development (17%), and employee experience (14%). Functions such as inclusion and diversity, ESG, and compliance remain at 2% or below, reflecting both appropriate caution and a lack of clear use cases. The pattern suggests HR teams are making targeted, pragmatic choices about where AI earns its place, rather than deploying it broadly across all functions.

Governance is the sector's most significant blind spot. Of organisations that have an AI policy, only one in four say it is clear and future-proof. More than half describe their policies as too restrictive and too tightly tied to specific current tools. Another 23% say their policies are too broad to guide practical behaviour. For HR professionals, where AI decisions can touch employment law, privacy, and discrimination risk, this is not an abstract concern. A basic governance framework covering approved use cases, human review requirements, data handling, and accountability does not need to be a lengthy document. The more important thing is to have one before scaling adoption.

The practical starting point for most HR teams sits in writing and summarisation: job descriptions, interview question banks, onboarding documents, and structured summaries of engagement survey free-text or exit interview transcripts. These tasks are high-volume, time-consuming, and low-risk, making them ideal first use cases. As the SHRM data makes clear, the majority of organisations haven't started yet. For HR professionals ready to move from intention to action, the window to build genuine AI capability before it becomes table stakes is still open, but it won't stay open indefinitely.