When a future nurse practitioner fails a high-stakes exam, it affects more than just their grade — it affects the pipeline in a healthcare system scrambling to meet the need for qualified clinicians. Exam preparation tools matter more than we typically acknowledge.
I've spent the last year thinking hard about what separates an AI study tool that actually moves the needle from one that students try once and abandon. My team at WGU Labs spent 120 days studying how 179 Family Nurse Practitioner (FNP) students used two AI-powered tools we designed (PathoPro Tutor, an adaptive quizzing app, and Clinical Companion, a patient simulation) to prepare for one of the most challenging objective assessments (OA) in the program in the course: Advanced Pathophysiology for the Advanced Practice Nurse.
There's a growing body of research connecting adaptive learning tools to improved outcomes, rooted in well-established principles like retrieval practice, spaced repetition, and immediate feedback — particularly in health professions education, where the stakes of exam readiness are high. We built PathoPro Tutor on those foundations, and wanted to know whether building on learning science actually translated to results in a real, high-stakes course.
92% of students who used PathoPro said they'd recommend it to a fellow student. Students' self-reported confidence across all six course competency areas climbed from 3.50 to 4.02 on a 5-point scale over the course of the study. 73% said they'd pay out of pocket for continued access.
But the most important things we learned weren't in the survey data. They were in the design decisions that got us there. Here are five principles that shaped how we built PathoPro Tutor, and what I'd tell my fellow learning experience designers building AI study tools today.
Five design principles for building AI study tools
1. Align the tool format to the assessment format
It doesn’t need to be exact, but it helps to come close. We designed PathoPro to generate multiple-choice questions that mirror the structure, language, and reasoning demands of the OA. Students repeatedly told us the study tool helped them get comfortable with the type of question, not just the content.
"The application just helped me get used to the layout and the type of questions on the OA."
If your tool doesn't look and feel like the thing students are preparing for, you are missing an opportunity to build their confidence in navigating the exam. That’s one less thing for them to worry about on test day.
2. Anchor questions to course competencies, not just general content
PathoPro questions map directly to the six course competency domains (the same framework students are assessed against). That alignment meant students weren't just reviewing general content; they were practicing the specific clinical reasoning the course required.
For designers, this means working closely with faculty and subject matter experts upfront. The time investment pays off in relevance — and results.
3. Design for formative practice, not just summative rehearsal
There's a meaningful distinction between formative and summative assessment experiences, and good learning design uses both intentionally. Formative interactions, like PathoPro's adaptive quizzing, prioritize detailed coaching: explaining not just the right answer, but why each wrong answer is wrong.
98% of students agreed that PathoPro's rationales were more effective for their learning than traditional study methods. As students build mastery, interactions can shift toward a more summative experience — withholding feedback until the end of a question set to simulate actual test conditions and build stamina.
4. Target knowledge gaps through adaptive depth
PathoPro tracks student performance and surfaces more questions in areas of weakness. This means a student who struggles with hematologic disorders gets more practice there, not another round of content they've already demonstrated they know.
94% of students said adaptive quizzing helped them identify gaps in their knowledge. The goal is efficient preparation: helping learners spend their limited study time where it matters most.
5. Let students steer
Students know where they're lost. PathoPro allows students to direct their own practice: selecting topics, requesting harder questions, and focusing on the areas where they feel least prepared. That sense of agency is more than good UX — it's good learning science. Learners who have control over their practice are more likely to engage deeply, return consistently, and build genuine fluency.
What's next?
We're continuing to test PathoPro Tutor in a follow-up study wrapping up this spring, in which a new cohort of FNP students can customize the frequency, depth, and tone of PathoPro's feedback. Early findings are already shaping how we think about personalization, and future studies will dig deeper into which feedback configurations actually shift learning outcomes. We'll follow the data wherever it leads.
