A framework for designing AI-integrated experiential learning experiences
As AI reshapes the skills and capacities graduates need, higher education faces increasing pressure to move beyond content delivery toward competency-based approaches that develop transferable, human-centered capabilities (Pellegrino & Hilton, 2012). Experiential learning provides the pedagogical foundation for this shift: research consistently shows that students build deeper understanding and durable skills when learning is anchored in real, meaningful doing rather than passive reception (Kolb, 1984; Morris, 2020).
The Experiential Learning in the Age of AI (ELAI) framework applies this logic directly to the AI era. Faculty begin by defining the competencies students need to embody, design an experiential learning structure that develops those competencies, and only then place AI intentionally within that structure. That order matters, because the effectiveness of AI in education depends heavily on instructional design and disciplinary context, and meaningful integration requires deliberate pedagogical decisions rather than broad adoption (Ahmed et al., 2025).
The result is a competency-driven approach to course design where AI serves the learning, not the other way around.
Each step comes with prompts you can paste into the AI tool you use (Bentley-approved tools are listed in the Guidelines for Responsible Use of AI). Fill in the brackets and treat the response as a first draft for your judgment, not a final answer.
Identify the core capacities students will embody by the end of the semester. These are the dispositions, skills and ways of thinking students carry forward beyond the course.
"Understands supply chains" is content. "Weighs competing evidence before deciding" is a capacity: it shows up in any course, job or conversation.
Select the model that best fits your course context, students and goals.
Students engage in extended, structured investigation of a complex real-world challenge, culminating in a tangible product or public artifact (Thomas, 2000).
A good fit when you want students to make something real for an audience beyond the classroom.
Students collaborate on an open-ended, ill-structured problem, applying cross-disciplinary knowledge and skills to develop viable solutions. The problem drives the learning (Savery, 2006).
A good fit when the messy problem itself is the teacher, and collaboration is central.
Students pose questions, investigate problems and construct knowledge through exploration, with or without a final product (Furtak et al., 2012; Minner, Levy & Century, 2010).
A good fit when curiosity and the process of investigating matter more than the deliverable.
Using your chosen model, design the step-by-step learning journey for your students. The Experiential Learning Design Toolkit offers scaffolded guidance for building out each phase.
Every capacity from Step 1 should show up in at least one phase. If one doesn't, the map isn't building it yet.
Review your learning map and make two deliberate decisions.
Identify moments where AI can deepen inquiry, support iteration, or extend what students can do. Consider, too, whether AI could take on non-essential tasks to create more time for meaningful practice, feedback and skill development.
Identify moments where using AI would bypass the core learning the experience is designed to create: the productive struggle, the human judgment, the creative risk.
Need ideas? The AI Pedagogy Strategy Deck has 14 ways to use AI that deepen learning, and Building Metacognition in the AI Era shows how to keep the thinking with students.
The Badavas Center is your thought partner for teaching projects of every shape and size, including designing an experiential course with ELAI from first sketch to launch.
Email the Badavas Center