Europe’s industrial backbone is facing a silent divide. While automation headlines are dominated by “lights-out” factories and billion-euro investments, the real test is unfolding on the shop floors of small and mid-sized manufacturers. For every multinational with a data-driven smart plant, there are dozens of European SMEs wrestling with legacy kit, skills shortages, and the ever-pressing demand to modernise.
Yet, the old storyline — that mid-caps are simply lagging — is now being upended.
Two firms, two futures
In Barcelona, textile machinery manufacturer CANMARTEX faced a classic industry dilemma: the high cost of defects in knitted fabrics. Waste rates of up to 10% during manufacturing were the norm, costing the company and its customers dearly. Recognising that this wasn’t just an operational headache but a competitiveness issue, CANMARTEX partnered with the DIH4CAT European Digital Innovation Hub and research centre Eurecat. The result was an AI-driven predictive quality control system — spun out as a new venture, Aracne — that analyses every stitch in real time using sensors, IoT, and photonics.
The outcome was transformative. Over a thousand sample analyses revealed more than 60% reduction in production defects. The new technology allowed CANMARTEX to raise machinery prices by 10%, boost sales by 20%, and create new skilled jobs. The project also secured nearly €92,000 in EDIH grants, validating the ecosystem model — one that blends internal investment, public funding, and deep technical partnerships.
“The innovation Eurecat and Canmartex have developed helps to make the industry more sustainable by ramping up production and cutting manufacturing costs while improving the innovation positioning of textile firms,” said Aracne CEO Enric Martí.
Far to the north, in coastal Finland, Innocode — a micro-enterprise of fewer than ten employees — tackled a different pain point: the unpredictability of machine failure. Traditional predictive maintenance solutions, dependent on extensive historical data, were out of reach for most SMEs.
Instead, Innocode, with the support of the Robocoast EDIH, developed the P100 system, a low-cost AI sensor that “listens” to factory equipment, creating a digital fingerprint of normal operations. When subtle changes arise, the system can flag issues weeks before a breakdown.
For a tiny firm, the support network was pivotal. Robocoast’s matchmaking provided business partners, testing environments, and even trade show access, saving Innocode an estimated 1–1.5 person-years in sales effort and significant direct costs. “Collaborative partners help our company to develop and grow,” CTO Mikko Kortelainen explained. “It would have taken years [to build these contacts] without the support from EDIH.” The result is not just better machine uptime and efficiency, but new market opportunities and regional job creation.
Sizing Europe’s AI adoption gap
Despite these examples, the gulf between large and mid-sized manufacturers remains. By 2024, only 21% of European mid-cap manufacturers and just 11% of small firms had deployed any AI tool, compared to 41% of large enterprises.
While Denmark and Sweden show adoption rates around 25–27%, much of the continent lags behind. And the urgency is rising: “The ‘time to react’ to change, leveraging AI to drive impactful change across their business and avoid falling behind competition, is only about 2 years now. The luxury of having 5-6 years to react to change is firmly over,” notes Adrian McGrath of UST, who emphasises that manufacturing firms may find themselves slipping behind without even realising it, as much of the transformation is “invisible” in daily operations.
Ask those working across the sector and the picture is clear: manufacturers are struggling to turn insight into action. Aaron Merkin, CTO at Fluke Reliability, says: “A lot of our customers have the right data. But their biggest challenge is turning the insights into meaningful action on the factory floor. Labour shortages are compounding this, forcing teams to deliver more with fewer resources, often under pressure to show rapid ROI from AI investments.”
Even where technology is available, “the gap between expectation and execution is widening.” He highlights that “our recent research found that while 97% of manufacturers plan to use AI to address the skilled labour gap, only 21% believe it will effectively reduce workloads and bridge the talent shortfall.”
Legacy infrastructure and fragmented systems are just as much of a challenge. Michael Thomas, executive director for digital at MARCH, sees how “poor data quality, legacy systems that were never built with AI in mind, and a workforce still learning to trust and engage with new technologies” all slow progress — especially in smaller companies. “Most factories already generate useful data through equipment like PLCs and OEM systems. The challenge lies in making that data accessible and usable.”
There’s often a misconception that becoming data driven requires a full-scale overhaul. In reality, “the challenge lies in making that data accessible and usable,” Thomas says, citing projects where simply connecting up legacy equipment and digitising data produced rapid gains.
For many SMEs, cost and lack of in-house expertise remain obstacles. “Without in-house data teams or the capacity to absorb risk, there is often a reluctance to commit to unfamiliar technologies,” says Thomas, who adds that network infrastructure — especially for remote sites — can be another “make or break” factor.
Claire Biggerstaff, manufacturing market lead for EMEA at Zebra Technologies, notes that “knowing where to begin, finding the right partners and resources, keeping pace with new technologies, and having appropriate goals and return on investment metrics” are all common pain points, while data quality and governance are fundamental. “SME and mid-cap factories need to make sure they have good data quality and management, and can remove obsolete, inaccurate, and duplicate data. Data owners and storage locations also need to be aligned and accessible for AI models to train and test on and access in real life.”
But ultimately, the human element is emerging as the most fragile link. “Even the most advanced analytics mean little if the workforce isn’t confident, contextually aware, and capable under pressure. The human factor is emerging as the most strategic and most fragile link in the value chain,” says Merkin. His data backs this up: “Our research showed 76% of industrial leaders acknowledge the importance of empowering their workforce to use these technologies; yet few have turned that into action. Upskilling alone isn’t enough. What’s needed is enablement at scale — a workforce that understands the ‘why,’ not just the ‘how.’”
Productivity, competitiveness, and the business case
What, then, does closing the gap deliver? As the CANMARTEX and Innocode stories show, measurable outcomes can be dramatic. At CANMARTEX, a >60% reduction in production defects and a 20% bump in machinery sales fundamentally changed the commercial trajectory of the business. According to McGrath, UST’s deployments of Vision AI across manufacturing clients have yielded “on average, 5% reduction in waste, 15% improved productivity, 20% reduced breakdown of equipment and 30% reduction in unscheduled stoppages.”
At Fluke Reliability, a building parts manufacturer achieved “more than $8.1 million in downtime cost savings and 637 hours of restored production, all within six months, as well as extending their asset lifespan by 10x,” using AI diagnostics and laser alignment tools. “We’ve seen across multiple AI deployments, success hinges on interoperability, usability, and measurable business outcomes,” says Merkin.
Similarly, Zebra’s case studies point to “a supplier to global automotive OEMs… that has improved and increased the production quality of electric battery caps” through machine vision, and other automotive OEMs “securing a 10–15% defect rate reduction in their quality inspection processes.” The bottom line: “the ability to deliver fast, actionable insights from day one reflects a broader principle we see across successful industrial AI deployments: time-to-value must be immediate, and the solution must be intuitive and scalable.”
Libra Speciality Chemicals: a cultural shift in action
For Libra Speciality Chemicals, a high-growth UK SME supplying global FMCG brands, digital transformation was about far more than new tools. The company faced mounting complexity, rising customer expectations, and the inefficiency of relying on manual, paper-based processes. “Manual methods were no longer enough,” Libra’s technical team recalls. Engineers could only spot issues when physically present on the line, while managers had to make decisions using delayed or outdated information. As production demands increased, the need for timely, accurate data became critical.
The turning point came with Project ARGUS, an industry initiative led by Procter & Gamble and backed by the Made Smarter programme. Libra became the manufacturing testbed, piloting new optical sensors capable of detecting microbial biofilms in real time — a game-changer for reducing the risk and downtime associated with contamination in chemical manufacturing. Traditional methods required samples to be incubated for up to three days, holding up production and tying up revenue in quarantined stock. The new sensors, integrated with XpertRule’s XpertFactory digital decision-intelligence platform, enabled real-time alerts and actionable insights, allowing teams to act before small issues could escalate.
Building on ARGUS, Libra rolled out XpertFactory across its core processes, transforming plant operations site-wide. Tank levels, once monitored manually, became digitised and visualised on intelligent dashboards, enabling procurement, sales, and supply chain teams to collaborate using the same live data. Live monitoring of agitator speeds, temperature, and power use unlocked further benefits, with trends tracked across batches to provide early warnings on inefficiencies or abnormal process conditions.
“Having centralised, real-time visibility has transformed how the site operates. Instead of reacting to problems, we anticipate and plan around them based on accurate, up-to-date information. It’s created a culture of visibility and transparency and everyone’s now pulling in the same direction,” says Dr Oliver Smith, technical manager at Libra.
The cultural impact was immediate: urgent customer requests could be handled faster, decisions became evidence-driven, and teams prioritised changes that delivered the greatest impact. “When a customer has an urgent request, our teams can check the status and respond quickly,” adds commercial director Sahd Hussain. “We’ve reduced delays and miscommunication because people no longer need to check updates; they just check the system.”
Looking ahead, Libra is extending digitalisation even further — eliminating paper trails, introducing digital sign-offs and automated compliance logs, and streamlining audit readiness in a regulated industry. “With XpertFactory, Libra has transformed its manufacturing operations from manual and reactive to real-time, intelligent and highly efficient. Operating from the same live data enables cross-functional collaboration and faster, more confident decision-making,” says Darren Falconer, technical director at XpertRule. “It’s not off-the-shelf; it’s something that evolves with us,” adds Hussain.
Libra’s journey stands as a model for how targeted digital investment, strong ecosystem partnerships, and a commitment to cultural change can deliver measurable business impact—turning digital transformation from a buzzword into the daily operating reality of a successful SME.
The foundations for Europe’s future
Across Europe, it’s the blend of funding, technical partnership, and peer support that enables most successful adopters to make the leap. CANMARTEX’s journey depended on public grants and innovation hub expertise. Libra’s transformation was accelerated through Project ARGUS and support from global partners. MARCH’s experience is that “partnerships can play a big role” — especially where vendors work side-by-side with clients to build internal confidence and run small-scale pilots before full rollout.
Building AI capability in manufacturing, everyone agrees, is not just about hiring data scientists or sending people on generic courses. “We’ve embedded AI skill-building into our apprenticeship and co-op programs. These aren’t theoretical exercises: they’re hands-on, real-world experiences designed to equip talent with the tools, context, and confidence to make an impact fast,” says Fluke’s Merkin.
MARCH adds that effective AI use “is about making sure the right skills exist throughout the organisation, especially among the engineers who are closest to the process.” Successful programmes, they argue, “start small, run pilot projects and make sure there is a clear operational benefit before expanding further.”
For Zebra, the Connected Factory Framework is about aligning the C-suite, IT and OT, “to create new ways of working that make everyday life better for their organisations, their employees and those they serve.” Their research found that “nearly three-quarters of manufacturing leaders… plan to reskill labour to enhance data and technology usage skills… Six in 10 leaders rank ongoing development, retraining/upskilling, and career path development to attract future talent as a high priority for their organisations.” Mobility-enabling technology and partnerships are seen as key.
What next? Beyond adoption to orchestration
A final theme is the move from isolated digital projects to AI “orchestrating” the entire production and supply chain environment. “A major AI theme in manufacturing is around adopting AI agents to take on a larger role in coordinating and controlling further automation across the manufacturing process and extended supply chain,” says UST’s McGrath. “These AI agents will evolve into an Agentic AI model, potentially disrupting many manufacturing processes and requiring very careful governance and responsible AI principles.”
The lesson from Europe’s fastest-movers is that adoption alone is not enough. The winners will be those able to connect the dots—technically, organisationally, and culturally — between real-world operational data, talent, partnership, and new models of automation. As Libra’s team puts it, “XpertRule listens. They don’t just offer a tool, they offer a partnership. It’s not off-the-shelf; it’s something that evolves with us.”
Perhaps the most significant lesson is that successful AI adoption in manufacturing is no longer a solitary endeavour. For every SME that makes the leap, there’s a network of funders, researchers, digital hubs, and peer firms behind them — proving that, in Europe’s industrial future, connectivity is as important as code.
