From Experimentation to Integration: Embedding GenAI in Business Higher Education through the Lens of Constructive

Synthesis: Zhou, Chai, Chilukuri, and Quach (2026) present a qualitative case study of generative AI (GenAI) adoption across 17 cases of business-module integration at a UK Russell Group university during the first year of implementation. Drawing on semi-structured interviews with academic staff spanning 24 undergraduate and postgraduate business modules, the study identifies three patterns of GenAI integration — ad hoc, blended, and constructive — distinguished less by frequency of use than by the degree of curriculum constructive alignment achieved. The central finding is that the balance between GenAI's pedagogical benefits and its associated risks was shaped by how deeply it was embedded in the curriculum: constructive integration cases reported stronger student engagement, capability development, and curriculum relevance, while ad hoc approaches proved more vulnerable to ethical concerns, overreliance, and inequality. The study extends Biggs' theory of constructive alignment to the GenAI context, showing that AI can be embedded within existing pedagogical strategies without requiring full curriculum redesign, and offers practical guidance for aligning GenAI with intended learning outcomes to support coherent, sustainable adoption in business higher education.

Key Findings

• Three patterns of GenAI integration emerged: Cross-case comparison of frequency, breadth, and pedagogical Scaffolding produced a typology of ad hoc (sporadic, unaligned use), blended (inconsistent incorporation across some activities or assessments), and constructive integration (GenAI systematically embedded across teaching, learning, and evaluative tasks with structured student guidance). Two of 17 cases were ad hoc, nine blended, and six constructive.
• Constructive integration yielded the strongest benefits: Cases with deeper, coherent embedding reported enhanced student engagement, capability development, curriculum relevance, confidence, and employability outcomes. Constructively integrated modules demonstrated a more favorable benefit-to-risk profile in the study's trade-off analysis.
• Ad hoc use amplified risks: Sporadic, unscaffolded GenAI use was more vulnerable to challenges including ethical and academic integrity concerns, student overreliance, inequality (e.g., access to premium tool versions), and ineffective or inappropriate outputs.
• GenAI enhanced existing pedagogies rather than replacing them: Successful integration commonly occurred within established frameworks — simulation-based, project-based, or experiential learning — positioning AI as a complementary enhancement and reducing the burden of curriculum redesign.
• External, industry-oriented motivation drove adoption: Business educators were strongly motivated by industry expectations and employability demands, extending prior research that emphasized internal motivations; disciplinary relevance shaped how readily educators viewed GenAI as a natural extension of their teaching.
• Institutional support was the critical enabler and barrier: A lack of structured training, clear policy guidance, and adequate infrastructure constrained educators from progressing beyond experimentation toward sustained, comprehensive integration, reinforcing the need for coordinated institutional investment.

Study Design & Method

• Qualitative case study of GenAI adoption at a UK Russell Group university during its first year of implementation, part of a funded project supporting staff integration of AI literacy into curricula.
• 17 academic staff participants (lecturers, teaching-focused academics, and program directors) selected via purposive sampling based on GenAI teaching experience and involvement in business-related modules; the unit of analysis was the individual educator and their pedagogical approach.
• 24 business modules spanning undergraduate and postgraduate levels (Levels 3–7) across areas such as strategic management, marketing, Sustainability, consumer behavior, financial reporting, and data analytics.
• Data collection: Semi-structured interviews via Microsoft Teams covering module context, implementation approaches, perceived impacts, barriers and enablers, and reflections.
• Analysis: Inductive thematic coding with parent themes (use of GenAI, enablers, barriers, benefits, challenges) and iterative cross-case comparison to develop the integration typology.

What this means for practice

• Instructors. Embed GenAI inside established pedagogies — simulations, projects, experiential learning — rather than pursuing full module redesign; the constructive cases reported the strongest engagement, confidence, and employability outcomes at the lowest workload barrier.
• Curriculum designers. Align GenAI use explicitly with intended learning outcomes, activities, and Assessment, since the ad hoc cases were the most exposed to ethical concerns and student overreliance.
• Instructors. Make evaluation the lever that shapes student use: the constructive cases required students to critically evaluate and co-create with AI in evaluative tasks, which separated purposeful use from sporadic use.
• Faculty developers. Provide structured training and clear policy guidance; the lack of both was the barrier that kept educators from moving beyond individual experimentation.
• Administrators. Fund the whole support ecosystem — policy, faculty development, and infrastructure — and close access gaps, because uneven access to premium tool versions drove inequality in the ad hoc cases.

Limitations

• The analysis rests solely on educator perspectives — 17 academic staff — so reported impacts on learning experiences and outcomes are second-hand and never measured from students.
• It is a single-institution case study at one UK Russell Group university during the first year of adoption, so findings may not be generalizable across institutional policy, digital readiness, or disciplinary norms.
• Participants were purposively sampled for existing GenAI teaching experience and their 24 modules, so the sample skews toward motivated early adopters.
• Given the rapid evolution of GenAI, the cross-sectional first-year data cannot show how integration strategies or institutional support change over time.

Citation

Zhou, X., Chai, Q., Chilukuri, B., & Quach, J. J. Y. (2026). From experimentation to integration: Embedding generative artificial intelligence in business higher education through the lens of constructive alignment. Journal of University Teaching and Learning Practice, 23(2).