Which Academic-Library Functions Are Using AI First?

p p p strongAcademic libraries are moving beyond experimentation with chatbots, but adoption is uneven: research support and data analysis are advancing faster than automated recommendations, tagging, digitization or facilities work./strong/p p Primary Research Group's survey of approximately 50 academic-library directors and other senior leaders found that 36% were using generative AI for research support and 32% for data analysis. Research universities moved even faster on data analysis, reaching 40%. Those figures put knowledge work&mdashnot building automation&mdashat the front of the current adoption curve./p p Other applications remain less common. Sixteen percent reported AI-supported personalized recommendations, and the same share reported chatbots. Image tagging stood at 12%, digitization tasks at 6% and predictive maintenance at just 2%. The pattern suggests that libraries are starting where staff can inspect outputs and where the work already involves summarizing, organizing or interpreting information./p p Budgeting has not caught up with use. Only 8% of respondents reported a specific budget for AI, with research universities dominating that small group. Without a visible budget line, pilots can multiply while responsibility for licensing, training, privacy review and evaluation remains scattered across departments./p p For library leaders, the next step is to turn isolated use cases into a portfolio. Record the task, tool, data involved, staff time saved, error risk and approval owner for every pilot. Then compare high-use applications with low-adoption ones to decide where formal investment is justified. The central management question is no longer whether libraries will use AI it is which functions can produce measurable value under library-grade standards./p p  /p/p/p