What we can learn from industrial, blue collar AI

Industrial companies, the ones that actually make stuff and keep the economy moving, are worth watching for how they will leverage AI. The industrial AI approach is decidedly practical, grounded in business reality and returns and designed to augment human labor that's in short supply. You could call it blue collar AI.

After spending a week at IFS Unleashed 2026, this blue collar AI theme is worth watching. Industrial AI is one of the reasons to be optimistic about enterprise AI even as bubble fears dominate AI infrastructure chatter.

Here's a look at some of the AI themes from industrial companies.

The opportunity. IFS CEO Mark Moffat noted during his keynote that 40% of the data for industrial companies isn't used. As a result, industrial operations are only running at about 60% of their potential and that reality means more unplanned downtime, productivity gaps and expertise that goes unused.

Moffat added that frontline workers only spend a third of their time with a tool in hand. The rest of the time is waiting, searching and filling in forms.

Add it up and the industrial sector is one area where AI, process and domain knowledge can converge and drive returns.

• IFS launches IFS Intelligence, Industrial AI Harness, eyes industrial productivity

Business case is priority because planning cycles are short and capital isn't infinite. Stratus CEO Bill Hayden said AI investments are selected based on return on investment, complexity and change management. Stratus, which builds and maintains signs and interior branding for large national brands, needs to think through the business case and timing because the planning cycles are compressed.

The company, an early customer of IFS Intelligence, now revolves around short planning cycles. "We're now thinking in like three months, 12 months max," said Hayden. "I have no idea what five years looks like."

Real returns. Multiple companies outlined returns based on money and time saved. William Grant & Sons, a distillery, cut a report migration from 200 days to 30 minutes. EightGroup cut the time to conduct an electrical inspection from 25 minutes to 60 seconds. Kodiak Gas’ Material Replenisher digital worker, which identifies parts for an asset, locates it in inventory and procures it if needed, showed an 85% efficiency gain for transactions processed via the AI agent.

• Kodiak Gas Services needs to scale, humans and digital workers are the answer
• How aviation maintenance looks at AI systems
• Palihapitiya on AI valuations, energy, auditors
• How IFS’ industrial customers are approaching AI use cases

Preventative maintenance was frequently cited as a use case for AI. "We need to move more from a reactive to a proactive and a predictive mindset, and getting a holistic approach of assets is critical in our journey. Second one is spare parts. Getting the right spare part at the right site at the right level is critical to operations. Too much access ties down of working capital, and less spare parts causes stockouts, causes downtime," said Chetan Balsara, CEO of Covia Corp.

Industrial AI use cases were also refreshing to this scribe who is frequently pelted with buzzword-laden slides, vaporware, new models that are .0001 better on a largely meaningless benchmark and questionable capital deployment.

Chamath Palihapitiya, CEO of Social Captial | 8090, said: ""I don't think that we have seen the investment returns that people have expected for the amount of money that people have spent in AI. It's a tale of two worlds. It's a lot of forward investment, speculatively building unbelievable enterprise value, but we don't know if that's really sustainable. And on the other side, real companies trying to figure out how to do it in a proper, responsible way."

The returns outlined at IFS Unleashed were from form-heavy workflows: procurement, parts sourcing, inspections, data fixes and training. Those projects went live in weeks, not years. One metric that was tied to field service was first-time fixes.

AI isn't replacing labor as much as it is making up for a shortage. Multiple IFS customers said AI use cases revolved training incoming workers replacing retiring technicians and offloading tasks that sap productivity.

Lauren Purcell, aircraft engineering manager at UPS, said the industry doesn't have the labor capacity to support the rate it's going. "In terms of workforce, we just can't keep up with the training to replace those that are leaving the industry and that are retiring," said Purcell. "We're seeing retirements from these people that have been in the industry for so long and have years and years of knowledge just walk out the door, and we're working to better the training processes and more automated training processes."

In a nutshell, industrial companies are pursuing people plus AI strategies. AI also has the potential to raise new workers to expert level faster.

Employee satisfaction and retention is an AI use case. MD Akil, director of project management at Kodiak Gas, said the company can lose workers for only a few extra dollars an hour. "AI can improve the user experience and improve employee satisfaction scores. One of the biggest problems that we have in our industry is retention," said Akil. "Having ease of usability at hand in house is a huge retention factor."

Retention is everything because the skilled labor shortage is real. Boeing expects the industry to need more than 700,000 maintenance technicians over the next 20 years with two-third of that sum replacing retirees.

Domain knowledge and industry context is everything. Industrial executives weren't focused on the latest, greatest models as much as applying their industry domain knowledge to AI. process is also critical. Building this context will be difficult given the number of people with institutional knowledge retiring.

"The most important thing is to reconstruct the organizational hierarchy of a company. I think the big opportunity is for folks to embrace and really understand how you get the tribal knowledge in people's heads encoded so that it's machine readable," Palihapitiya said.

Governance, safeguards and trust are in the forefront. Governance was a core topic among industrial companies. AI will require safeguards and observability at every step of a process. In industrial settings, you can't simply blame an AI agent for a bad outcome. Machines may take action, but you'll own the outcome.

Industrial companies realized that they couldn't have a human sign off on every work order, but will have to rank the risks. Human checks on automated systems for critical infrastructure make sense. Other processes will simply slow down processes that could be automated.

Industrial AI is a long game. Industrial customers at IFS weren't all on the cloud or the latest version of the platform. Some didn't have their data strategies down. Other customers were early adopters of AI models. While robots were plentiful on the show floor and around, most customers weren't thinking humanoid robots. Industrial AI will be built one use case at a time because capital isn't falling out of the sky.