- Share - [](https://www.facebook.com/sharer.php?u=https://about.ebsco.com/markdownify/node/163619) - []() - []() - []() # How to Prove Library Value as Research Moves into AI As AI changes how researchers find and use information, traditional usage metrics may tell only part of the story. Libraries need new ways to understand and demonstrate the value their resources provide. ![electric car AI rings](https://assets.ebsco.com/asset/1e28b64a-09f3-4b44-acc2-d1a004d4b926/EBSCOhost-AI-Exchange-How-to-prove-library-value-as-research-moves-into-AI-780.png) 26 October 2026 ## What You'll Learn - AI-mediated research may make library value less visible through traditional usage metrics. - Metrics developed around human clicks and full-text retrievals may need to evolve alongside research behavior. - Library leaders should examine whether their current measures capture how licensed information is creating value. - Publishers, libraries, and information providers need to work together as new patterns of access and measurement emerge. ## A Familiar Library Question Is Becoming Harder to Answer Library leaders have always needed to answer a deceptively simple question: Are our resources being used, and are they delivering value that justifies the investment? Usage statistics help answer that question. Full-text retrievals, searches, and other established measures give libraries tangible evidence they can use when evaluating collections, managing budgets, and communicating their contribution to the institution. That need is not going away. But the way researchers interact with information is changing. As AI becomes part of research workflows, a researcher may benefit from licensed scholarly information without following the same sequence of searches, clicks, and full-text retrievals that libraries have traditionally measured. If measurement continues to focus only on those established interactions, library leaders could be looking at an incomplete picture. That creates an important question for academic libraries: Are we still measuring the right thing? ## What Can Electric Cars Teach Us About Library Metrics? Consider miles per gallon. For a gasoline-powered car, MPG works well. Fuel goes into the car, the car travels a certain number of miles, and one familiar number describes its fuel efficiency. Then hybrid vehicles arrived. MPG still told us something, but it no longer told the whole story. Some of the vehicles’ travel was powered by electricity, so looking only at gasoline consumption missed part of what was happening. With a fully electric vehicle, the limitation becomes obvious. Its MPG is effectively zero, but that does not mean the car failed to move. It means MPG is no longer the appropriate measure of the vehicle's performance. The engine changed. The metric did not. Library usage may be approaching a similar transition. Traditional full-text retrievals and other COUNTER usage measures were designed around recognizable human interactions with information. They remain meaningful measures. But if AI-mediated research creates new ways for users to benefit from library-provided information, those metrics may eventually represent only part of the activity libraries need to understand. The lesson from the car analogy is not that the old measure suddenly becomes useless. It is that one measure may no longer be sufficient to describe the whole system. ![](/themes/custom/cog_ebsco/images/pullquote-icon.png) ### “The value of library-provided content does not disappear just because the way researchers access it changes. As AI reshapes research workflows, the way we measure that value needs to evolve, too.” Bonnie Leavitt Vice President, Product Management EBSCO Information Services ## Why Does This Matter for Library Leaders? For library deans, this is more than a reporting issue. Library leaders are expected to connect investments in collections and services to broader institutional priorities. They need to demonstrate that the library contributes to research, teaching, learning, and the university's mission, often while managing significant financial pressure. If research behavior changes faster than the metrics used to describe it, that job becomes more difficult. A decline in one familiar usage measure could have several possible explanations. It could indicate lower demand. It could reflect a change in the research experience. Or some activity could be happening through pathways that existing measures were not designed to capture. That distinction matters when libraries make renewal decisions, discuss budgets with university leadership or evaluate how well their collections are supporting scholarship. A metric is useful only when library leaders understand what it measures and, equally important, what it does not. ## What Questions Should Libraries Be Asking Now? There is not yet one new metric that can replace the measures libraries rely on today. A more useful starting point is to examine whether the evidence used to demonstrate value still reflects the way researchers work. Library leaders can begin by asking: - **Where are researchers beginning their work?** Are research workflows expanding beyond traditional library interfaces? - **How are licensed resources contributing to those workflows?** Can the library identify where its investments are supporting research even when the interaction looks different? - **What activity is visible today, and what may not be?** Which parts of the research journey are represented in existing reports? - **Are collection decisions relying too heavily on a single indicator?** Could multiple signals provide a more complete view of value? - **What information will university leadership need?** How can the library connect resource use with research and institutional priorities rather than reporting activity alone? These questions do not diminish the importance of established usage standards. They help put those standards in context as research behavior evolves. ## How Can Libraries Prepare for a New Measurement Model? The immediate task is not to discard existing measures. It is to broaden the conversation about what library value looks like. Libraries can start by documenting how researchers are using AI within their institutions, identifying where current reporting provides good visibility, and noting where gaps are emerging. They can also bring this question into conversations with publishers and information providers. If AI-mediated discovery and access become a larger part of scholarly research, libraries will need meaningful ways to understand how licensed information contributes to those experiences. This is inherently a shared challenge. Libraries need evidence to make responsible collection decisions. Publishers need appropriate ways to understand the use of their content. Information providers need to consider how new research experiences can preserve the visibility and accountability libraries rely on. EBSCO Information Services is examining these questions as part of its work to understand how AI is changing research behavior and library-centered research workflows. That work is consistent with EBSCO's broader approach to AI, which emphasizes collaboration with librarians and researchers rather than treating AI as separate from the academic community it serves. ## The Metric Hasn't Changed. The Research Engine Has. The shift from gasoline-powered to electric vehicles required new ways to measure performance. AI may create a similar need for academic libraries. Traditional usage measures remain valuable, but they may not tell the complete story of how library resources contribute to research. For library leaders, the question is expanding beyond "How many times was this resource accessed?" to "How did this resource contribute to research?" Answering that question will be increasingly important as libraries demonstrate their value to the institutions they serve. ### How Is AI Changing Research? Learn how libraries and universities are exploring the role of AI in research and discovery. [ Learn more ](https://about.ebsco.com/ebscohost-ai-exchange/universities) Category: [Technology ](/blogs/ebscopost/category/technology) Tags: [AI](/blogs/ebscopost/tag/ai) [Academic libraries](/blogs/ebscopost/tag/academic-libraries) [Research databases](/blogs/ebscopost/tag/research-databases) - Share - [](https://www.facebook.com/sharer.php?u=https://about.ebsco.com/markdownify/node/163619) - []() - []() - []() ## Related Posts [ ![hands typing on a keyboard](https://assets.ebsco.com/asset/6486e417-5930-4a04-adae-b0b18f9a4ad8/EBSCOhost-AI-Exchange-Your-students-are-already-using-AI-780.png "hands typing on a keyboard") ](/blogs/ebscopost/your-students-are-using-ai-what-information-shapes-their-answers) [Your Students Are Already Using AI. What Information Is Shaping Their Answers?](/blogs/ebscopost/your-students-are-using-ai-what-information-shapes-their-answers) 5 October 2026 [ ![](https://assets.ebsco.com/asset/cb8fd56b-dd98-4d08-bb43-a0fa1da96311/ebscohost-ai-exchange-blog-780.jpg) ](/blogs/ebscopost/ebscohost-ai-exchange) [The Future of AI Depends on What Powers It: Introducing EBSCOhost AI Exchange](/blogs/ebscopost/ebscohost-ai-exchange) 26 May 2026 [ ![Photo of books with overlay and triangles](https://assets.ebsco.com/asset/8467013f-5594-49e1-9cf6-2b900e78c18e/folio-5-ways-to-future-proof-blog-image-780.png) ](/blogs/ebscopost/5-ways-libraries-are-building-future-ready-services) [5 Ways Libraries Are Building Future-Ready Services](/blogs/ebscopost/5-ways-libraries-are-building-future-ready-services) 6 April 2026