Earlier in September, 2024, Andreessen Horowitz published the article “Every White-Collar Role Will Have An AI Copilot. Then An AI Agent.” I fully agree, Angela Strange and James da Costa, thanks a lot for the publication – since at remberg we see it the same way. The automation potential for white-collar jobs is huge, also in the industrial segment. Jobs like technical helpdesk, dispatchers, planners, and machine programmers are going to be impacted. However, what about the blue-collar workforce? We still have a massive labor shortage here.
Over the past few months, we looked closely into this, and we can confidently say that just as every white-collar role will have an AI copilot, so will every blue-collar role. Obviously, these roles are more difficult to fully automate. So we will see fewer AI agents, but even in the industrial world, many tasks are repetitive and prone to AI automation.
The question is, how can we use AI to have an industrial workforce with less or even no limits? Let’s explore how AI agents and copilots will handle industrial tasks, plan, reason and take action in the future.
What are blue-collar roles?
A blue-collar role typically refers to jobs that involve manual labor, skilled trades, or physical work. These jobs are usually hands-on, involving manufacturing, maintenance, or service work, and may require specialized training or apprenticeship. Blue-collar workers are often paid hourly wages and may work in industrial settings or fields requiring physical exertion.
Example industries and typical blue-collar jobs
Manufacturing
Machine operators, welders, quality control inspectors etc.
Energy & Utilities
Line workers, power plant operators, gas fitters etc.
Telecommunications
Cable infrastructure installers, network technicians, tower climbers etc.
Machinery & Plant Engineering
Machine technicians, service specialists etc.
Buildings & Facility Management
Building maintenance technicians, HVAC installers, electricians etc.
IT & Electronics
IT hardware installers and technicians etc.
Food & Beverage
Production line workers, quality control inspectors, maintenance staff etc.
Hospitality
Housekeepers, maintenance staff, kitchen workers etc.
Healthcare
Maintenance technicians, biomedical equipment technicians, janitors etc.
Medical Devices
Field service engineers, calibration technicians, equipment installers etc.
Vehicles & Fleet Management
Mechanics, fleet maintenance workers, vehicle inspectors etc.
Construction
Electricians, carpenters, heavy equipment operators etc.
Field Services
Field technicians, service engineers, maintenance workers etc.
Logistics & Transportation
Truck drivers, warehouse workers, forklift operators etc.
Public Sector
Public works technicians, building maintenance workers, utility workers etc.
No news: Industrial labor shortage is still a thing
We can take a look at the numbers, but it’s not new that we have a huge blue-collar workforce shortage. Quick reminder at this point, according to government data from the U.S. Census Bureau and the Bureau of Labor Statistics, we are facing 2.1 million unfilled jobs in the U.S. and about another 2 million in the EU in manufacturing alone, not to mention logistics, transportation and all the other sectors mentioned above.
Challenges in Attracting Younger Generations to Blue-Collar Roles
I assume many blue-collar jobs seem less attractive to younger generations due to a combination of societal biases, the push toward college education, and the appeal of more flexible careers. Compared to working in physically demanding and often remote environments like flying with a helicopter somewhere to fix an oil platform where you cannot easily post anything on Instagram, in comparison to working from your couch in your pajamas is somewhat outside of the comfort zone here.
With the prosperous western society we live in, I don’t see excited younger generations lining up for blue-collar roles anytime soon, so we need to intervene. Of course a massive migration of skilled workers is a big lever, but also not getting easier with the political shift to the far right that we see in many countries, I don’t think that’s going to solve the shortage of blue-collar workforce anytime soon either.
Leveraging AI to Address Industrial Labor Shortages
Of course, as entrepreneurial techno-optimists, we are motivated to now apply the latest wave of AI technology to solve some of the world’s biggest challenges, one of which being the industrial labor shortage. If we can’t fully automate blue-collar jobs with AI agents, we should empower them, increase their productivity and automate as much as we can.
We analyzed Gartner’s key use cases impacted by Generative AI from the white-collar space and derived the most impactful use cases where AI software is eating industrial labor.
The three most impactful and low-hanging fruit AI use cases to get to an industrial workforce with no limits
- AI Content Discovery in the Industrial World
Search & Knowledge Management: With technological advancements impacting the physical world, blue-collar roles also face a rapidly changing environment—perhaps not as fast as white-collar jobs, but still significant. According to McKinsey, the average worker spends 19% of their time searching for information. While this statistic likely applies more to white-collar workers, it’s reasonable to assume a similar figure for blue-collar workers. If two-thirds of this time can be attributed to blue-collar workers, and we reduce it by 90% through AI content discovery & search, then it results in time savings equivalent to a 10% productivity boost. In the industrial sector, assets are the single common denominator of all jobs – the machines, tools, and equipment that workers use, operate, maintain, or rely on to do their jobs. We will drive significant impact here if we can help blue-collar workers to always access the right knowledge and information around their assets in seconds. So besides blue- and white-collar roles, also every asset is going to have a copilot.
Note: We will see an additional boost when LLMs are better in multi-modality, e.g. when they can interpret circuit diagrams of industrial assets. We believe that’s just a matter of time until the billions of EUR invested in pre-training the next foundational models will yield these capabilities. Until then we can “lean back” and automate all the things we describe here – enough to do and time is on our side.
Analysis, Diagnosis & Troubleshooting: When an asset fails, AI presents a strong opportunity to apply the shift-left principle, enabling workers to troubleshoot and resolve issues independently without relying on support from the asset’s OEM or external providers. This can significantly reduce downtime, though the impact will vary depending on the specific work setup and environment. Here you can see a video of an AI copilot for assets in action.
- AI Content Creation in the Industrial World
Language Generation:
In blue-collar jobs, documentation is essential after a task or work order completion, whether it’s reporting on tasks, hours, or parts used, and in many companies, this is also required due to compliance regulations to ensure high-quality products and adherence to industry standards. AI can dramatically accelerate this process, especially in field service, installation, maintenance and repair roles, by allowing technicians to simply type or speak a few notes into their mobile devices, with generative AI completing the report. This is speeding up documentation by 10x. When combined with an API-first vertical application layer, it is easy to capture data that then allows integration into legacy ERP systems for transactional follow-on workflows like invoicing. Currently, technicians spend up to 20% of their time on documentation. If AI can reduce this by 50%, we unlock another 10% more blue-collar capacity globally. Alternatively, if you want to think about it in ROI, considering a typical hourly cost of €60 for a technician, this productivity boost could return approximately > 10k € savings per technician per year.
Code Generation:
Industrial assets, even the less complex ones now increasingly rely on embedded code (software has indeed eaten the world) and control units, meaning field service and maintenance technicians need to become more adept with this aspect of their work. Just as GitHub Copilot has proven successful for developers, a copilot for technicians will help them interact with machine code more efficiently. Siemens has already launched the Industrial Copilot for this use case. Although specific savings data is not yet available, GitHub reports up to a 55% faster completion rate in coding tasks, which hints at the potential for similar gains in industrial settings.
Image/Video Generation:
AI-powered image and video generation tools like Synthesia will transform industrial worker training by creating realistic, customizable training videos with minimal effort. As assets become more and more complex, traditional training methods struggle to keep up. Synthetic videos can simulate intricate tasks, demonstrate advanced machinery operation, or outline detailed safety protocols, offering workers a clear understanding of new technologies. These AI-generated videos can be tailored to specific equipment or workflows, ensuring that workers receive highly relevant, up-to-date training.
Audio Generation:
AI-powered audio generation can bring significant advantages to industrial settings by enabling technicians to interact with assets directly through voice. Imagine a technician calling an AI agent to ‘talk’ about a machine, where the AI provides help on maintenance, operations or troubleshooting steps. This hands-free communication increases efficiency by allowing workers to access critical information while remaining focused on their tasks or when they have no stable internet connection to chat with an AI copilot, taking a phone call might work.
- AI Conversations in the Industrial World
A copilot for assets e.g. is a conversational AI that then is a lever to drive points 1 and 2 in AI Content Discovery and Creation. Although Gartner puts this as a separate category I am not really in line with the MECEness. What they probably mean is the fact that you can “talk” to something, e.g. a machine. By integrating virtual assistants and natural language interfaces like AI copilots, blue-collar workers can quickly summarize long work order information received from their white-collar colleagues. This allows them to understand tasks in the field more efficiently, access the history of previous work, and identify who handled the job before.
Additionally, many blue-collar workers operate in environments where the local language is not their own. AI-powered translation tools can bridge this gap, supporting workers with documentation in the language although the complex documentation is in a different one. Field service technicians traveling abroad can also rely on AI to translate reports and other communications for local customers.
Quoted from the a16z article and transferred to the vertical in which we are operating:
In the industrial sector, just like in white-collar environments, “the most natural place for these copilots and agents to integrate is within existing workflows or systems of record.” This includes platforms such as ERP, CMMS, EAM, and FSM solutions. These systems of record (SOR) are where copilots (and later, agents) can perform specialized tasks. It’s also the logical place for any new user interface to reside, such as prompts for the AI agent.
Since industrial incumbent systems are legacy, to make AI live, we need a modern system of record & action for assets
The difference to other sectors is that the existing systems of record in the blue-collar industrial environment are much more legacy than those in the white-collar environment. For example, according to the Gartner EAM Report 2024, IBM still has 64% of its Enterprise Asset Management installed base on-premise.
That obviously opens up a big opportunity for digital challengers when the incumbent system of record is NOT the most natural place for copilots and agents to appear. This is a vertical where there is room for a new system of record, context, and action for assets, the common denominator for blue-collar work.
With the remberg Asset Platform, we are building exactly this modern, multimodal, AI-first system of record & action which will be the #1 place for industrial AI copilots and agents to live.
Summary: AI will shift labor into software also in the industrial sector
While (industrial) white-collar jobs may be easier to automate with the immediate, high-impact use cases mentioned by Gartner, the potential for blue-collar productivity gains through AI is very significant. Even focusing on low-hanging AI enhancements, this will unlock a 20-25% productivity boost across various blue-collar roles. For example, the U.S. has approximately 6 million technicians working in installation, service, maintenance, and repair, with another estimated 7 million in the EU. If AI could free up 20% of their capacity, we are looking at the equivalent of gaining an additional 2.6 million technicians, significantly increasing asset uptime, and advancing key initiatives like renewable energy deployment or infrastructure improvements.
This results in substantial economic implications, with a potential of over 100 billion EUR in value —not to mention the “opportunity costs” of additional value created by unlocking these productivity gains. And that’s not just for blue-collar workers.
In the future, industrial workers, AI copilots and agents will work closely together using the next generation of Asset Platforms. At remberg, our mission is to take the lead in enabling an industrial workforce with no limits.
If you want to get going or have ideas for further AI use cases in our industry, feel free to reach out directly.
Back