SMEs struggle with AI adoption due to skills gaps, data issues and strategy failures

A newly published study sheds light on the complex factors impeding the adoption of artificial intelligence (AI) among small and medium-sized enterprises (SMEs), offering a structured roadmap to accelerate responsible and effective integration. The research, titled “Artificial Intelligence Adoption in SMEs: Survey Based on TOE–DOI Framework, Primary Methodology and Challenges”, was published in Applied Sciences in June 2025. 

The study critically examines the organizational, technological, and environmental conditions shaping AI uptake in SMEs, while proposing solutions grounded in empirical and conceptual analysis.

What are the key challenges to AI adoption in SMEs?

Despite AI’s transformative potential, the study highlights that SMEs struggle with AI implementation due to interrelated challenges spanning internal capabilities, system compatibility, financial limitations, and regulatory complexity. Through the Technology–Organization–Environment (TOE) framework, complemented by elements from the Diffusion of Innovations (DOI) theory, the researchers identify ten major adoption barriers that SMEs encounter.

Technologically, SMEs are hindered by poor data quality, limited IT infrastructure, and the complexity of AI tools. Many firms also lack access to generative AI (Gen-AI) models due to financial or knowledge gaps. Organizational challenges include severe shortages in AI talent, cultural resistance to change, and the absence of structured methodologies to guide adoption. In particular, the fear of job displacement and skepticism toward automation hinder alignment between human resources and AI systems.

From an environmental perspective, insufficient collaboration with public institutions and limited access to shared innovation platforms further exacerbate SME constraints. In many cases, SMEs operate in regulatory grey zones, with unclear guidance on responsible AI practices, leading to underinvestment and cautious experimentation.

The study emphasizes that these challenges are systemic and interdependent, and overcoming them requires holistic, rather than isolated, interventions. For example, merely purchasing AI software is ineffective unless cultural, strategic, and technical alignment is simultaneously pursued.

How does the study propose to support AI adoption in SMEs?

To bridge the gap between theoretical potential and practical implementation, the study introduces a six-phase roadmap designed to guide SMEs through the adoption lifecycle of AI. This methodology reflects the TOE–DOI framework and addresses both perceptual and structural barriers:

1. Readiness Assessment: Evaluate infrastructure, skills, leadership alignment, and financial capacity.
2. Strategy Definition: Identify business-relevant use cases aligned with measurable goals.
3. Tool Selection: Choose scalable, compatible, and cost-effective AI technologies.
4. Pilot Deployment: Test solutions through low-risk pilot projects and measure performance.
5. Training and Upskilling: Build internal capabilities to reduce dependence on external consultants.
6. Monitoring and Adjustment: Continuously evaluate AI integration outcomes and adjust processes accordingly.

This structured methodology encourages SMEs to embed AI into their strategic core, rather than treating it as a peripheral experiment. The researchers also promote the adoption of modular, cloud-based, and open-weight AI systems, such as LLaMA, DeepSeek-R1, Mistral, and Falcon, as cost-effective alternatives to proprietary models. These solutions allow SMEs to retain control over their data while reducing implementation complexity and licensing fees.

Moreover, the roadmap is designed to be context-sensitive. It accounts for SMEs’ informal decision-making structures, lean operational models, and often limited digital maturity. In doing so, it offers a realistic, phased path forward, rather than an idealized transformation blueprint.

What broader implications does the study reveal for policy and practice?

The research concludes with a call to action for policymakers, ecosystem actors, and SME leaders to adopt inclusive, responsible, and scalable AI strategies. A major contribution of the study is its emphasis on responsible AI governance - a facet that is frequently overlooked in SME settings. The authors argue that embedding ethical considerations such as transparency, data protection, and stakeholder involvement into AI planning is no longer optional. Lightweight governance mechanisms, explainable AI interfaces, and employee engagement processes are suggested as practical ways for SMEs to futureproof their innovation journeys.

In addition, the study stresses the importance of building public–private innovation ecosystems that can provide shared infrastructure, technical guidance, and financial incentives tailored for SME adoption needs. Evidence from multiple regions, including the EU, India, and Ghana, validates the utility of the TOE–DOI framework across different sectoral and national contexts, enhancing the generalizability of the findings.

The authors recommend further empirical research to validate the proposed roadmap and refine its application across diverse industries. This includes field testing, participatory workshops, and comparative case studies that can assess real-world effectiveness.