Webinar: Digital & AI Deployment within Existing Companies

A webinar hosted by Sisak-Moslavina County (Croatia) was held on May 26, 2026 on behalf of the Community of Practice for Industrial Modernization and AI – Human-Centric Digital & AI Transformation. The session, which drew approximately 80 participants, highlighted practical pathways for SMEs to adopt AI, drawing on case studies from Croatia and Japan as well as cross-regional expert panel insights. There was a consensus that AI value hinges on high-quality data, integration into real workflows, and human-centric change management. Another shared view was that SMEs require modular, fast-impact solutions and ecosystem support to progress from pilot to production. Last but not least, speakers and panelists repeatedly emphasized a shared commitment to responsible AI, stressing the importance of human-in-the-loop approaches for high-impact decisions.

The meeting began with opening remarks from Andreja Šeperac, Regional Coordinator of Sisak-Moslavina County, and Hidefumi Imura, Country Coordinator for Japan, IURC.

A keynote speech by Igor Buzov, Head of the AI Center Lipik (Croatia), showed that financial challenges and time constraints were the largest barriers to AI deployment in regional SMEs. The center addresses limited budgets and a skills gap through evening, practice-oriented AI training and European Digital Innovation Hub (EDIH) support. Its SME strategy focuses on small, modular AI solutions that deliver quick wins and build momentum for broader digital transformation. Another recurring success factor was human-centric change management to address fear, time constraints, and workers’ confidence in their capabilities.

Entrepreneur and CEO of Grow, Josipa Bencek (Croatia), shared her experience working with clients to develop advanced digital solutions and immersive, interactive experiences. She stressed the need to move beyond simply offering or installing code toward building operational, scalable AI systems that are integrated into existing production flows together with the client. Example deployments included an autonomous AI assistant that turns knowledge into action and a GIS/computer-vision-aided detection service to identify hidden real estate properties and improve municipal revenue.

In Japan, Aichi Prefecture runs subsidized trials that help SMEs select and adopt appropriate digital tools. These tools are offered for testing free of charge. One successful example involved an SME specializing in the manufacture of automotive bolts that was seeking a way to facilitate skills transfer from veteran to younger workers. Using digital tools “Dive”, the company created smart glasses that detected and recorded veteran workflows and used generative AI to automatically transcribe speech into structured text, resulting in an educational manual for younger workers.

Hiroshima Prefecture (Japan) showed how an industry–academia–government ecosystem can scale talent and applied R&D through large training volumes, joint projects, and shared infrastructure. One of the core features introduced was the Hiroshima University Digital Manufacturing Education & Research Center’s company-funded, two-year employee dispatch to Hiroshima University since 2022 (5 participants; 3 returned as visiting researchers). Due to its success in promoting exchange between industry and academia, the program is set to expand, with additional companies becoming eligible in the future.

The three presentations were followed by a panel discussion moderated by Domagoj Crnkovic, an educational specialist at the PISMO gamedev business incubator of Sisak-Moslavina County (Croatia). The discussion featured timely topics on how to address AI amid gaps in data quality and ethical challenges. Silvia Benati, in charge of EDIH ER2Digit at ART-ER, Emilia-Romagna (Italy), shared results from digital maturity assessments of 92 firms in the region, which showed that companies still require foundational investment and human resource development for AI adoption. Meritxell Bassolas Ribas Cataloni, Director of the Innovation Unit at the Computer Vision Center (CVC), said that one reason companies are unsuccessful with AI deployment is that they tend to lack strategy and adopt a technology-first mindset—without first framing the business problem or while underestimating the complexity of integrating with legacy systems. According to Mylène Richard of the Regional Digital Campus in Auvergne-Rhône-Alpes (France), regional ecosystems (digital campuses, competitiveness clusters, technical centers, AI institutes, EDIHs) are essential to support governance, training, and secure transitions from proof of concept (PoC) to production.

Panelists largely characterized data quality and data foundations as the primary gating factor for SME AI adoption, citing inconsistent structures, incomplete records, siloed systems, and constrained machine interfaces.Panagiotis Aivaliotis, Executive Director of the Teaching Factory Competence Center (TF-CC) in Western Greece (Greece), agreed on these data limitations and suggested that AI could help mitigate them—for example, by generating synthetic data via digital twins and physics-based simulation. Lastly, the panel framed ethics as operational risk management, especially in high-impact domains (public administration, health/life sciences, and socioeconomic decisions) where bias, privacy exposure, and loss of trust have outsized consequences. There was a joint consensus that human in the loop approach was essential for safe Ai deployment.

The event closed with comments from Oscar Pratt, Business and Academia Expert for the IURC-AA, and Andreja Šeperac’s suggestions for joint collaboration in the future.