Making Assessment Matter

Becky Croxton and Megan Oakleaf

AI as Assessment Partner

Practical Applications and Critical Cautions, Part 2

Becky Croxton is strategic assessment librarian at Colorado State University Libraries, email: becky.croxton@colostate.edu. Megan Oakleaf is associate dean for academic affairs and professor at the Syracuse University iSchool, email: moakleaf@syr.edu.

© 2026 Becky Croxton and Megan Oakleaf

Making Assessment Matter is a five-part C&RL News series focused on maximizing the impact of academic library assessment. The first article introduces strategies for launching assessment projects designed for action and impact. The second explores how centering participants and stakeholders increases the likelihood that assessment results will be used. The third article identifies common pathways for translating assessment results into decisions and action, and the fourth focuses on communicating evidence in ways that resonate with key audiences and support meaningful change. Part 2 of this fifth and final article explores how AI can be incorporated thoughtfully into each stage of the assessment cycle without displacing practitioner expertise. Together, the series equips librarians to use assessment as a driver of learning, decision-making, and continuous improvement.

Introduction

Academic library assessment practitioners have much to consider, starting with the end in mind to ensure assessments lead to action; centering participants to generate meaningful, collaborative evidence; anticipating pathways that connect findings to decisions; and communicating results in ways that reach the right audiences and inspire change. Each of these practices requires knowledge, skill, and professional judgment. Into this established and evolving practice, artificial intelligence (AI) has become impossible to ignore. Part 1 of this article (C&RL News, July/August 2026) addressed two foundational questions: What guardrails are necessary to ensure that AI supports, rather than undermines, trustworthy user-centered assessment practice and what institutional opportunities and risks should practitioners keep in mind as they incorporate AI into their assessment work? This second part, grounded in ALA’s Guidance on the Use of Artificial Intelligence in Libraries,¹ ACRL’s recently approved AI Competencies for Academic Library Workers,^2, and ACRL’s Proficiencies for Assessment in Academic Libraries,³ takes up the remaining question: In what ways does it make sense to incorporate AI into library assessment practices across the assessment cycle?

Practical Applications across the Assessment Cycle

Many conversations about AI in assessment focus on what an AI tool can generate, analyze, or summarize. Perhaps a more productive framing is to consider how AI can support specific moments in the assessment cycle, an approach that aligns with the Proficiencies for Assessment in Academic Libraries, which emphasize intentional, reflective, and action-oriented assessment practice.

Rather than treating AI as a general-purpose solution, this section situates AI use within five key phases of the assessment cycle: (1) choosing focus and setting goals; (2) gathering and analyzing data; (3) making decisions; (4) sharing results and implementing change; (5) revisiting goals; and beginning the assessment cycle again, as illustrated in Figure 1.⁴ Organizing AI-supported assessment in this way keeps the work grounded in purpose, prevents distraction, and reinforces the foundational role of assessment: “to learn from the assessment and use that learning to make decisions or take actions that lead to positive impacts.”⁵ Across the assessment cycle, AI is most effective when it is leveraged as a “human in the loop”⁶ thinking partner to improve efficiency and accelerate thinking,⁷ surface patterns or gaps worth investigating, and support communication while leaving framing, interpretation, and decisions firmly in human hands.

1. Choose Focus and Set Goals: Using AI to Support Purposeful Planning

An important first step when planning an assessment is to “start at the end,” with attention to the decisions that will be made and the actions they are meant to support. As outlined in the first article in this series, one practical way to do this is by expressing an assessment’s purpose as user stories that clarify who needs to direct or make a change, what question or need can be acted upon, and why the change matters.⁸ AI can serve as a thinking partner at this stage to help generate and refine user stories for consideration. When guided by carefully designed prompts, AI tools can draft multiple user stories in a specified format, allowing assessment teams to explore alternative framings, different stakeholders or action‐takers, and possible impacts and to test whether proposed questions clearly connect evidence to action. The sample AI prompt in Box 1 demonstrates how a carefully designed prompt structure, language, and constraints shape the relevance and usefulness of AI output.⁹ AI-generated user stories function as starting points that practitioners can then refine to ensure alignment with local context and decision-making needs.

AI can also support early planning by helping practitioners scan and synthesize institutional context. For example, AI can be used to review strategic plans, accreditation reports, or committee materials to surface recurring priorities, language, or expectations that an assessment might address. Used in tandem with user stories, this approach can help ensure that assessment purposes are not only actionable for stakeholders but also well aligned with broader institutional goals while leaving judgment about relevance and priority in human hands.

Clarifying purpose in this way naturally surfaces who needs to be involved in the assessment. During the planning and goal-setting phase, AI tools can help brainstorm possible stakeholders and draft a preliminary stakeholder register that identifies possible roles and responsibilities, levels of influence and interest, values, expectations, and potential impact, as described in the second article of this series.¹⁰ Assessment teams can then review, validate, and refine this draft, using AI to reduce startup effort while retaining human judgment about whose voices shape both the assessment and the actions that will follow.

Note: Box 1 presents a fully drafted AI prompt to demonstrate how assessment practitioners can use AI as a thinking partner when clarifying purpose and stakeholders. In subsequent phases of the assessment cycle addressed in this article, the goal is not to repeat full AI prompts but to highlight prompt parameters: the recurring structures, constraints, and types of questions that shape responsible AI use at different moments in assessment work. As such, the remaining boxes summarize key prompt components rather than reproducing complete scripts, emphasizing how AI’s role intentionally shifts across the cycle while professional judgment remains central.

2. Gather and Analyze Data: Using AI to Enhance Design and Discovery

Once an assessment’s purpose and stakeholders are clear, the next step in the assessment cycle is to determine what evidence to gather and how to make sense of it, emphasizing intentional design, appropriate methods, and critical engagement with data.¹¹ AI can support this work by extending the capacity of assessment practitioners and helping them to do the following:

• Brainstorm meaningful data to collect
• Assess the scope of the data being considered and the degree to which it will answer the user stories or questions at hand
• Identify gaps in draft data-gathering and analysis plans

In the data collection and design phase, AI can assist practitioners by drafting survey, focus group, or interview questions and aligning them with stated goals. AI may also be used to propose potential comparison groups or institutional peers based on institutional characteristics, offering a starting point for benchmarking discussions. In such cases, AI helps reduce startup effort while maintaining methodological rigor.

AI can be particularly useful during data analysis, especially when the volume or complexity of data makes fully manual approaches impractical. For example, AI may assist with text analysis of open-ended survey responses or chat transcripts, exploratory review of longitudinal usage data, or reflection on what may not be captured in the dataset but could matter for demonstrating value in a given area. AI can also suggest analytic approaches or raise questions about alternative ways to examine the evidence.

In these situations, AI functions as a thinking partner, not a decision maker, that can surface potential trends or anomalies for human review. Practitioners must set clear analytic parameters, examine how analyses were conducted, and consider how data quality, context, and algorithmic bias may shape results and disproportionately affect different populations, as illustrated in the AI draft prompt criteria in Box 2. AI-generated outputs should be treated as signals that prompt further inquiry rather than as conclusions to be accepted at face value. There is no substitute for human judgment in determining whether patterns are meaningful, misleading, or actionable.

At this point in the cycle, interpretation focuses on sense-making: determining what the evidence suggests, where uncertainty remains, and clarifying the implications and limits of observed patterns.¹² Interpreting what the findings mean for decisions and action belongs to the next phase of the assessment cycle.

3. Make Decisions: Using AI to Inform Evidence-Based Choices

Assessment matters most when results inform decisions and lead to action. The third article in this series emphasized the importance of anticipating “pathways to impact”—the concrete ways that assessment results can be used to guide change, from updating policies and reallocating resources to shaping strategic plans, fostering innovation, or even deciding not to act.¹³ Anticipating these actions helps ensure that assessment findings are positioned to support decision-making rather than stall at the reporting stage.

During this phase of the assessment cycle, AI can help practitioners think more expansively and prepare more intentionally for the conversations, decisions, and actions that assessment is meant to inform. Acting as a thinking partner, AI can assist by drafting alternative decision scenarios or surfacing implications for different stakeholder groups, as illustrated in the sample prompt criteria in Box 3. In this way, AI supports, rather than substitutes for, the deliberate human work of turning evidence into informed, values-aligned decisions.

4. Share Results and Implement Change: Using AI to Facilitate Communication and Action

Assessment results achieve impact when they are shared in ways that support understanding, decision-making, and follow-through. As discussed in the fourth article in this series, communicating assessment findings is not a final step but rather a strategic practice that connects evidence to action by reaching the right audiences with clear, relevant, and actionable messages.¹⁴ Effective communication gives voice to the data, supports collaboration and partnership, and helps ensure that assessment insights inform implementation and demonstrate outcomes, impact, and value rather than remaining siloed in reports.¹⁵

AI can support this phase of the assessment cycle by assisting with synthesis, translation, and efficiency while leaving message framing and interpretation in human hands. When practitioners have already identified their goals, key messages, and proof points, AI tools can help draft assessment reports using defined parameters, generate executive summaries, and extract key takeaways from lengthy reports. These outputs should be treated as drafts; practitioners must always review and verify AI-generated content to avoid misrepresentation, overemphasis, and loss of nuance.

AI can also support this phase by helping assessment teams coordinate how findings are communicated across audiences. For example, AI can assist in preparing multiple versions of the same findings for different audiences, building on earlier planning tools like stakeholder communication matrices that clarify who needs results and how they should be shared.¹⁶ In this capacity, AI helps streamline information sharing, reduce redundancies in drafting efforts, and support collaboration across roles and units¹⁷ while leaving decisions about emphasis, framing, and recommendations for action to practitioners. The AI prompt criteria in Box 4 illustrate how AI can facilitate coordinated communication in practice.

Finally, AI can support routine communication tasks, such as generating first drafts of routine reports or suggesting data visualizations to clarify complex findings. In this phase of the assessment cycle, AI assists with the execution of communication and implementation while humans remain responsible for decision-making and action-taking.

5. Revisit Goals and Begin the Cycle Again: Using AI to Close the Loop

Assessment does not end with implementation. After decisions are made and actions taken, the assessment cycle calls for reflection on what was learned, what changed, and whether original goals, questions, and assumptions require revision for subsequent cycles of assessment. AI can support reflective practice in this phase by identifying gaps in assessment efforts, suggesting adjustments to assessment processes, and refining assessment documentation for the next cycle.

AI across the Assessment Cycle

AI is most useful when it accelerates thinking, expands capacity, and reveals opportunities for insight, without displacing professional judgment. Human expertise remains essential for setting purpose, interpreting meaning, and ensuring that assessment contributes to positive change rather than generating noise. Organizing AI use around the assessment cycle can help practitioners stay grounded in why, not just how, they assess. When used as a thinking partner, AI can help libraries do assessment work that is not only more efficient but also more intentional and actionable.

Conclusion

AI is not an author, an interpreter, a library or institutional expert, or an accountable decision-maker. In library assessment, AI is most valuable when it functions as a thinking partner that augments, accelerates, and supports the work of humans and least valuable (or even harmful) when it displaces the professional judgment, contextual knowledge, and ethical grounding that assessment practitioners bring to their work. Situating AI use within the assessment cycle keeps purpose and professional judgment, rather than technology, at the center. Every AI output requires careful review, verification against what practitioners know about their data and their communities, and consideration of whose perspectives are represented and whose may be missing. When that rigor is brought to AI-supported assessment work, libraries are better positioned, as discussed in Part 1, to contribute data and perspectives to broader institutional analyses, thereby strengthening understandings that are incomplete without them.

As an assessment partner, AI is only as effective as the practitioners who direct it, evaluate its outputs, and ensure it serves library and institutional communities responsibly. Although the future of AI remains uncertain, one thing seems clear: In careful, knowledgeable hands, AI-supported assessment can be a genuine partner in achieving long-standing goals by bringing users and colleagues together, adding insights to deepen understanding, informing decision-making, and ensuring that assessment leads to positive change for library communities.

Notes

1. American Library Association (ALA), “Guidance on the Use of Artificial Intelligence in Libraries,” approved June 2026, https://www.ala.org/tools/standards-and-guidelines/guidance-use-artificial-intelligence-libraries.

2. Association of College & Research Libraries (ACRL), “AI Competencies for Academic Library Workers,” approved October 2025, https://www.ala.org/acrl/standards/ai.

3. Association of College & Research Libraries (ACRL), “Proficiencies for Assessment in Academic Libraries,” revised and approved June 2023, https://www.ala.org/acrl/standards/assessment_proficiencies.

4. Peggy L. Maki, “Developing an Assessment Plan to Learn about Student Learning,” Journal of Academic Librarianship 28, no. 1 (2002): 8–13, https://doi.org/10.1016/S0099-1333(01)00295-6; Rebecca Croxton and Megan Oakleaf, “Chapter 6. Assessment and Evaluation,” in Academic Libraries: An Introduction, ed. Michael Crumpton and Nora J. Bird (In Press).

5. Becky Croxton and Megan Oakleaf, “Pathways to Impact: Anticipating Action in Library Assessment,” College and Research Libraries News 87, no. 2 (2026): 79–83, https://doi.org/10.5860/crln.87.2.79

6. ALA, “Guidance on AI Use - Library Values and AI, 1. Public Good.”

7. ACRL, “AI Competencies – 4.1 Apply AI for Task Efficiency and Quality Enhancement.”

8. Megan Oakleaf and Becky Croxton, “Start at the End: Strategies for Actionable Assessment Results,” College and Research Libraries News 86, no. 9 (2025): 382–85, https://doi.org/10.5860/crln.86.9.%25p; Megan Oakleaf, Library Integration in Institutional Learning Analytics, EDUCAUSE, 2018, https://library.educause.edu/resources/2018/11/library-integration-in-institutional-learning-analytics.

9. ACRL, “AI Competencies – 4.3. Develop Effective Prompting Strategies for Optimal AI Output.”

10. Megan Oakleaf and Becky Croxton, “From Subjects to Partners: Centering Participants in Library Assessment,” College and Research Libraries News 86, no. 10 (2025): 449–54, https://doi.org/10.5860/crln.86.11.449.

11. ACRL, “Proficiencies for Assessment – 3. Designing, Collecting, and Analyzing.”

12. ACRL, “Proficiencies for Assessment – 4. Reflecting and Making Meaning from Results.”

13. Becky Croxton and Megan Oakleaf, “Pathways to Impact: Anticipating Action in Library Assessment,” College and Research Libraries News 87, no. 2 (2026): 79–83, https://doi.org/10.5860/crln.87.2.79.

14. Becky Croxton and Megan Oakleaf, “Evidence to Action: Communicating for Meaningful Change,” College and Research Libraries News 87, no 4 (2026): 168–73, https://doi.org/10.5860/crln.87.4.168.

15. ACRL, “Proficiencies for Assessment – 5. Communicating and Taking Action.”

16. Oakleaf and Croxton, “From Subjects to Partners: Centering Participants in Library Assessment.”

17. ACRL, “AI Competencies – 4.2 Use AI to Facilitate Communication and Collaboration in the Workplace.”