Factories don’t run on machines alone: the $1 Trillion skills gap in manufacturing

This article is part of Brighteye’s Heavy Industry Series, a three-part exploration of how startups are transforming the physical economy. 

In this first piece, part one, we look at manufacturing’s often-overlooked workforce layer. Part two dives into construction’s project-based memory loss. And part three will explore digital transformation and its implications on people in the mining industry.

-----

Factories are evolving, but without a workforce enablement layer, the machines won’t matter. Let’s fix that before production grinds to a halt.

Machines and operations are getting a full upgrade, becoming more capable and arguably more independent than ever before. However, the people inside them are being left behind and this is having serious productivity consequences:

• New tools are being rolled out faster than employees can fully absorb. With inadequate upskilling on new machinery and changing workflows, employees are struggling to keep up. Upskilling typically happens off the manufacturing floor, in classroom-based settings, making it difficult to apply new knowledge in the flow of work.

• Experienced operators are retiring, taking critical know-how with them. Due to poor systems for upskilling and knowledge capture, factories face a generational gap in operational know-how. When seasoned technicians leave, the institutional knowledge built over decades goes with them.

• New hires are being dropped into complex, high-stakes environments with little more than a rushed walkthrough and a binder of outdated training documents. Training is often ad-hoc and insufficient, putting workers and factory performance at risk. Poor onboarding also drives dissatisfaction in a sector already struggling to attract talent - proper training helps team members feel valued.

• Inefficient use of data science. As highlighted by Maximillian Hahnenkamp, co-founder of Scavenger AI, accessing and analysing data still requires IT or coding expertise, which is often lacking in manufacturing settings, meaning that decisions often aren’t being taken based on real, operational data.

While expectations of productivity and learning outputs rise, support systems haven’t kept pace. Training is still delivered through PowerPoints and PDFs. Context is missing, guidance is scarce, and most tools feel bolted on, as opposed to within the flow of work on the manufacturing floor.

Further to this, the skills gap is widening.

According to Deloitte, 2.1 million U.S. manufacturing roles may go unfilled by 2030, potentially costing the economy up to $1 trillion. Europe’s trajectory mirrors this - according to The Manufacturer, 75% of manufacturers with operations in the UK identify skills shortages as their biggest barrier to growth, followed by recruitment challenges (36%) and talent retention (32%). Aging workforces, increasingly complex systems, and a lack of scalable infrastructure to keep skills current and visible are other key issues.

We don’t only have a labour shortage - we have an enablement gap. There’s a disconnect between what workers are equipped to do and what they are being asked to do. This is slowing down the transformation that industrial automation promised. According to Jean-Baptiste Ronssin, founder of Yoshu, this is the ‘very essence of the problem that manufacurers are facing’.

As explained by Jasper Admiraal, founder of Baseboard there are two key mechanisms driving this shortage. Firstly, innovations are happening in very quick succession, and secondly machines are getting increasingly complex. This means that operators need to be highly skilled or well instructed to manufacture high-tech and high-quality products rapidly. However, managing the production and assembly processes of complex machines with short time to market is time-consuming and prone to errors. This means that operators have to work with either outdated documentation, or no documentation at all.

Why now? Macro & workforce trends

• Aging workforce & knowledge drain: As stated in the introduction, a significant proportion of highly skilled workers are retiring without effective knowledge transfer systems in place.

• Regulatory complexity: Compliance is becoming more demanding, from safety protocols to sustainability requirements, requiring tools that help frontline workers stay current without slowing production.

• On-the-job enablement > classroom learning: Manufacturers need real-time, contextualised learning tools that embed training into daily workflows, not siloed LMS systems.

• Automation ≠ deskilling: As robotics and software proliferate, workers need more technical training—not less. From digital twin interaction to robot calibration, the bar is rising, and systems must support that rise.

The new stack: From hiring → training → real-time enablement

A new generation of startups is stepping in, not with course libraries, but with workflow-native tools that embed support, visibility, and knowledge right where the work happens.

At present, most manufacturers still rely on fragmented, outdated systems: spreadsheets to track skills, classroom sessions for training, and binders of SOPs for support. This outdated stack fails to meet the moment. It slows onboarding, obscures skills visibility, and puts safety, quality, and efficiency at risk.

The new stack puts enablement in the flow of work. It enables real-time troubleshooting, contextualised training, seamless knowledge transfer, and live compliance tracking, in most cases, directly from the manufacturing floor.

By embedding intelligent safety systems into the manufacturing floor, manufacturers reduce downtime, reinforce trust, and keep experienced operators on the job, making safety a driver of productivity, not just a compliance metric.

These aren’t just “learning” tools, they are workflow-native enablement layers. They provide help where and when the work is happening, not buried in an LMS that is discovered away from the core parts of manufacturing roles. It’s onboarding that doesn’t start with slides and binders but with a smart assistant guiding you moment-to-moment, reducing time-to-productivity. These systems amplify the workforce we already have, helping them stay safe, competent, and hopefully enjoying their work.

Real Impact: Company Examples

• Yoshu: Delivers AI-powered, real-time answers to frontline teams, surfacing operational knowledge when and where it’s needed to prevent errors and support better decision-making. Manufacturers using Yoshu have seen:

  • 30% reduction in onboarding time within six months

  • 10–15% improvement in quality KPIs

  • Up to 50% reduction in unplanned downtime

  • Reduction in time spent locating critical documentation with real-time knowledge assistants - ROI exceeding 400% in the first year alone

• Scavenger AI: Uses spatial intelligence and computer vision to optimise material flow, layouts, and real-time process visibility. Employees are able to ask Scavenger business questions in plain language and get instant, actionable answers

  • One German manufacturer reduced production data analysis time by 70% using Scavenger’s AI-powered tools.

• Baseboard: Provides structured, collaborative documentation for operations, helping teams align around processes, scale best practices, and improve execution across sites. They automatically update the production documentation so that everyone works with the most recent data.

  • Reduce work planner time by 50% and scrap and rework significantly.

Conclusion

We’ve upgraded our machinery. Now it’s time to upgrade the human support systems around them.

Despite growing automation, factories don’t run on machines alone, they run on people with the knowledge, confidence, and tools to adapt and improve.

A new generation of startups is rising to meet this need, building the infrastructure for a more capable, efficient, and resilient industrial workforce.

In the medium term, we believe that people will remain indispensable. As Secretary of Commerce Howard Lutnick recently noted, even the most automated factories still need skilled workers to install, repair, and maintain complex machinery (Fortune). The real shift will be toward technical trade roles, where the ability to work alongside advanced robotics and AI will be critical.

But in the long term, there’s a bigger risk to consider: some factory roles will likely disappear or evolve as automation matures. Workforce enablement, therefore, isn’t just about short-term productivity gains; it’s about preparing people for what’s next, whether that’s a more technical role on the shop floor or a transition into new industries. The factories that invest in this now won’t just be more efficient today; they’ll build a future-ready workforce for tomorrow.

Part 2, on construction, coming soon!