D2L, a global leader in learning innovation, today released Beyond the Hype: How Embedded AI Can Help Higher Education Move from Experimentation to Instructional Impact, a new report drawing on research conducted by Digital Promise on behalf of, and with support from D2L. The research examines how faculty, staff and instructional leaders are using D2L Lumi in higher education settings, where they are finding value and what institutions may need to support meaningful AI adoption.
The Digital Promise research-practice partnership examined the experiences of higher education faculty, staff and instructional leaders. The findings show that while participating institutions were broadly supportive of AI adoption, educators responded most positively when AI was connected to practical, clearly defined instructional tasks and integrated into the environments where they already work.
Among participants using D2L Lumi, creating assessment questions and providing feedback to students were the most frequently cited use cases. D2L Lumi Quiz emerged as a particularly strong early use case, with participants citing its ability to generate multiple questions quickly while still allowing educators to review and approve the output before use.
The research found:
- Practical use cases can help build confidence in AI. Participants responded positively to AI that addressed clearly defined instructional tasks, with D2L Lumi Quiz described as straightforward, efficient and useful.
- Embedding AI into existing workflows can reduce friction. Participants valued being able to use D2L Lumi within D2L Brightspace rather than moving course materials between the learning environment and standalone AI tools.
- Course context can contribute to trust. Participants connected D2L Lumi’s use of materials within D2L Brightspace with greater confidence in its outputs, while only a small proportion identified bias or inaccuracy as a challenge compared with their experiences using other AI tools.
- Educators want AI to support their work rather than replace it. Participants saw value in using AI for routine and repetitive tasks that could create more time for course design, student interaction, mentorship and other high-value work requiring educator judgment.
- Instructional designers can play an important role in adoption. Participants frequently described instructional designers as important resources for introducing faculty to AI and connecting the technology to immediate instructional needs.
- Educators want AI to be evaluated on more than adoption. Participants identified efficiency, effectiveness and accessibility as important criteria for assessing whether AI supports teaching and learning.
“What this research makes clear is that meaningful AI adoption is not driven by what the technology can produce. It is shaped by what educators can confidently stand behind,” said Dr. Cristi Ford, Chief Learning Officer at D2L. “Educators saw the greatest value when AI fit naturally into their work and allowed them to inspect, refine and remain accountable for the output. Saving time is important, but it is not the measure of success. The real opportunity is to use AI to create better conditions for teaching and learning while preserving the judgment, creativity and connection that educators bring to their students.”
The research also identifies areas institutions and technology providers should continue to address as AI adoption develops. Participants expressed interest in greater customization, richer contextual capabilities and more flexibility in how AI can draw from instructional resources. They also raised considerations around AI-generated feedback, including the importance of preserving the role of instructor reviews which help educators understand student progress.
The findings suggest that institutions looking to expand AI use should pair technology with targeted support for educators. Participants described one-on-one assistance from instructional designers as particularly useful because it helped faculty understand how AI could address a specific instructional need at the moment it arose.
The report also identifies efficiency, effectiveness and accessibility as a potential framework for evaluating educational AI. This includes whether AI reduces instructor time and effort, produces useful outputs, supports student learning and contributes to inclusive, high-quality learning experiences.
Read Also: Reframe Systems Raises $40M to Industrialize Homebuilding and Expand Microfactory Network
