Why the future of maintenance depends on women - and AI that learns from them

After 25 years in enterprise asset management, I've learned this: maintenance and reliability teams are at their best when they combine hard-won experience with fresh thinking. Right now, our industry faces a very real skills shortage as seasoned technicians retire and fewer people enter the trades. But we also face a quieter contradiction - one that we can fix. While organizations struggle to fill critical roles, we're still overlooking a huge portion of the available talent pool. Women remain dramatically underrepresented in maintenance roles, even as demand for skilled workers keeps rising.

The data makes the urgency clear. In Ultimo's Maintenance Trend Report, 63 percent of organizations identify an aging workforce as the most critical trend, and 50 percent report major disruption from recruitment challenges. At the same time, women represent only 7.6 percent of manufacturing maintenance technicians. That gap isn't just unfair, it's inefficient. In a labor crisis, leaving talent on the sidelines is a business risk we can't afford.

The encouraging news is that momentum is building. I see it in the energy of organizations like Women in Reliability and Asset Management (WIRAM), which we proudly sponsor. WIRAM brings together an inspiring mix of professionals: technicians who began on the shop floor, administrators who moved into technical roles, engineers, and PhDs specializing in reliability and energy systems. They represent different industries, countries, and career paths, but they share something important: proof that talent has never been the constraint. Access, visibility, and opportunity have been.

We're also at a once-in-a-generation inflection point because workforce transformation is converging with artificial intelligence. Maintenance organizations are beginning to deploy agentic AI - systems that operate like digital coworkers. They learn from day-to-day interactions, capture decisions and outcomes, and build a kind of institutional memory that doesn't retire or change shifts. Done well, this technology can preserve tacit knowledge: the "feel" of a machine that's about to fail, the subtle warning signs that don't show up on a dashboard, and the troubleshooting instincts that come only with time.

That's a big deal. It means junior technicians can ramp faster with guidance that reflects proven practices. It means administrative load can shrink as AI supports routine reporting and documentation, freeing teams to focus on hands-on work that truly requires human judgment. It means expertise becomes scalable - available when and where it's needed.

But there's a catch, and it's one we can address if we act intentionally: whose expertise are we capturing?

AI isn't neutral. It learns patterns from the people and processes it observes. If our maintenance workforce remains overwhelmingly male while these systems are being trained, we risk encoding a narrow set of perspectives - problem-solving approaches, communication norms, and assumptions about what "expertise" looks like. We've already seen how bias can show up in AI systems when training inputs are skewed. Maintenance and reliability can - and should - do better.

This is where I'm genuinely optimistic. Because the solution is practical, measurable, and aligned with performance.

First, organizations need to recruit and retain women in maintenance roles as a business imperative. That includes modernizing job branding, widening sourcing channels, ensuring equitable shift and career progression opportunities, and creating team environments where respect and belonging are non-negotiable.

Second, we should strengthen pathways. Partnerships with schools, STEM programs, community colleges, and apprenticeship networks can make maintenance careers visible earlier - especially to young women who may never have been encouraged to consider the field.

Third, as we implement AI, we must design knowledge capture intentionally. Who is interviewed? Whose work orders become "gold standard" examples? Who validates recommendations? Diverse teams should be part of every step because the quality of the system depends on the breadth of the expertise it learns from.

Finally, we need to elevate the women already doing this work. They should be recognized as subject matter experts, mentors, and leaders whose knowledge will shape both the next generation of technicians and the digital tools that support them.

The window is open right now. AI systems are learning from today's workforce and today's practices. If we broaden participation today, we build smarter systems tomorrow systems that reflect the full spectrum of human expertise and make maintenance careers more accessible for everyone.

The people we need to meet tomorrow's reliability challenges already exist. Many of them are women ready to contribute, lead, and innovate. If we pair that talent with intentional AI deployment, we don't just solve a staffing problem, we strengthen the future of maintenance itself.