Navigating institutional and capability barriers in agentic artificial intelligence adoption: evidence from small and medium enterprises in Bangladesh

Преодоление институциональных барьеров и барьеров, связанных с возможностями, при внедрении агентного искусственного интеллекта: данные о малых и средних предприятиях Бангладеш
Md Asadul Islam, Mahfuzur Rahman, Francesca Dal Mas, Shamim Ehsanul Haque, Umme Hani
2026-02-12

BangladeshSME adoptionagentic artificial intelligencedynamic capabilitiesinstitutional barriers
Purpose This study aims to explore the barriers and solutions associated with the adoption of Agentic artificial intelligence (AAI) among small and medium-sized enterprises (SMEs) in Bangladesh. AAI, an advanced evolution of generative AI, offers potential for enhanced knowledge sharing, autonomy and organizational performance, yet its adoption in developing economies remains poorly understood. Drawing upon institutional and dynamic capabilities theories, this study investigates how contextual, structural and capability-related factors influence AAI adoption. Design/methodology/approach A qualitative research design was used, involving 14 semi-structured interviews with SME owners, technology experts and government officials in Dhaka, Bangladesh. Thematic analysis was used to identify key barriers and solutions, supported by triangulation, peer debriefing and member checking to enhance credibility. Findings The analysis identified eight primary barriers: slow internet speeds, limited knowledge and skills, inadequate infrastructure, lack of government support, political instability, high adoption costs and absence of a legal framework. To overcome these challenges, participants proposed developing awareness and education programs, building technological infrastructure, improving internet connectivity, establishing supportive legal and policy frameworks, offering financial incentives, fostering public–private partnerships and integrating AAI education into academic curricula. These findings reveal that AAI adoption is shaped not only by firm-level readiness but also by institutional and ecosystem maturity. Originality/value To the best of the authors’ knowledge, this research is among the first to examine AAI adoption in SMEs within a developing economy empirically. By combining institutional and dynamic capabilities theories, this study contributes to the growing literature on intelligent systems, offering a holistic framework that links institutional context, organizational capability and knowledge management to AAI adoption in emerging markets.
1
AAI adoption is constrained by institutional conditions such as insufficient government support and political instability, beyond firm-level technological readiness.
2
Integrating agentic AI education into academic curricula is proposed to strengthen the skills needed for SME adoption.
3
Participants recommend awareness and education programs, improved infrastructure and connectivity, supportive policies, financial incentives, and public–private partnerships.
4
The findings indicate that successful AAI adoption depends on both organizational capabilities and broader institutional and ecosystem maturity.
5
The study identifies eight primary barriers to agentic AI adoption among Bangladeshi SMEs, including slow internet, inadequate infrastructure, limited skills, high costs, and absent legal frameworks.

Agentic artificial intelligence adoption in small and medium-sized enterprises in Bangladesh

Institutional, structural and capability-related barriers and proposed solutions shaping Agentic artificial intelligence adoption

Publication Details
Publication Date
2026-02-12
Journal
VINE Journal of Information and Knowledge Management Systems
Publisher
Emerald Publishing Limited
ISSN
2059-5891
Cited by
21
Access Type
Author Information
Authors
Md Asadul Islam
Mahfuzur Rahman
Francesca Dal Mas
Shamim Ehsanul Haque
Umme Hani
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%