The Technician Shortage Is a Knowledge Problem. AI Is Being Asked to Spread It

Most plants describe their technician problem as a hiring problem: too few applicants, too many retirements, too long to bring a new starter up to speed. Two pieces of research published this year suggest the hiring gap is real but only half the story. The other half sits inside the maintenance teams plants already have, where the difference between the best technician on a shift and the rest of the crew is costing more than most operations leaders have measured.

The Demand Curve Has Moved Away From the Line Itself

A study released on September 10 by Deloitte and the Manufacturing Institute, the workforce affiliate of the National Association of Manufacturers, projects that manufacturing technician employment could grow six times faster than employment in production occupations between 2025 and 2030. Across manufacturing and adjacent-industry technician roles, employers may need to fill 2.3 million openings in that period, a figure that includes replacement demand from retirements and other exits as well as growth. Nearly half a million manufacturing technician roles are open now.

The full Deloitte Insights analysis shows how far the shift has already run. More than 4.5 million people worked in manufacturing and adjacent technician occupations in 2025, and employment of industrial machinery mechanics in manufacturing rose 47% between 2010 and 2025. The more a plant automates, the more of its labour moves from running equipment to keeping it running.

The study's proposed answer is to widen the pool. It identifies nearly 2 million technicians in adjacent industries whose skills may transfer, and finds that more than two-thirds of them work in construction, retail trade, other services, wholesale trade and government. Its central argument is that AI embedded in everyday workflows could deliver the plant-specific knowledge, skills and guidance those recruits lack, so that a capable electrician from a building services firm becomes productive on a packaging line faster than traditional training allows.

The Gap Inside the Team Costs More Than the Gap in the Roster

Bringing in more people only helps if the knowledge they need can reach them. Aquant's 2026 Field Service KPI Benchmark Report, built from anonymised data on nearly 30 million service events, 7 million assets and $8.3 billion of service costs across 161 service organisations over three years, shows what happens when it does not.

The industry first-time-fix rate sits at 77%. Top performers reach 88% and bottom performers 60%. Median time to resolution is 4.5 days, against 2.5 days at the top and 10 days at the bottom. Failed visits, the return trips and repeat call-outs that follow a fix that did not hold, account for 25% of total service cost at the median, 14% at the best organisations and 44% at the worst.

The most telling number is the skills gap, which Aquant defines as the difference in first-time-fix rate between an organisation's top technicians and everyone else. In best-in-class teams it is 2.9 percentage points. In underperforming teams it is 10 points. In other words, the strongest organisations do not simply employ better technicians. They have found ways to make their best people's judgement available to the whole crew, so that the result of a call depends far less on who answered it.

Those teams also keep their people. Aquant found top companies retain 87% of employees, against 66% at underperformers. A technician who keeps getting sent to jobs without the context to finish them is a technician more likely to leave, which deepens the shortage the plant was trying to hire its way out of.

What AI Can and Cannot Carry

Aquant's data points to where AI-assisted guidance is most likely to pay back. It estimates that one in five cases could be resolved remotely, and that sending a technician when one is not needed raises the cost impact from 1.4% to 18.3%. It puts the potential service cost reduction typically in the low teens, reaching 26% in some segments. Those savings come from triage, diagnosis and fix history reaching the person on the job before they reach for their tools.

Deloitte is careful about the limits. Its analysis cites the firm's 2026 Global Human Capital Trends survey, in which 85% of respondents said building the ability to adapt was important and only 7% said they were making great progress. The study frames AI as one part of a broader investment in skills, trust and workforce partnerships rather than a substitute for them. Funding is following that view: in April, Google.org committed $10 million to the Manufacturing Institute to build AI skills in the manufacturing workforce through its FAME programme.

The practical constraint is the one plants already know. An AI assistant can only spread knowledge that has been captured, and in most maintenance departments the most valuable fixes live in the heads of a few senior technicians, on paper work orders, or in free-text fields nobody has structured.

What Operations Leaders Should Do Now

• Measure the gap inside the team: compare first-time-fix and repeat-visit rates for your top technicians against the rest of the crew, by asset type, before deciding how many more people you need.
• Price the failed visits: total the labour, downtime and parts cost of every repeat call-out over the last year, and track it as its own line rather than burying it in general maintenance.
• Capture fixes as structured data: require digital job records that log the symptom, cause, parts and steps taken, so the reasoning behind a good repair outlives the shift that made it.
• Recruit from adjacent trades deliberately: map which of your technician roles draw on skills held by electricians, HVAC and building services technicians, and build plant-specific onboarding around the difference.
• Pilot AI guidance on your most repeated faults: start where the history is richest and the fix is most often missed, and judge the pilot on first-time-fix rate rather than on usage.
• Pair new starters with captured expertise: give recruits from outside manufacturing access to both a mentor and the plant's fix history from their first week, not after they have learned the hard way.

The technician shortage will not be solved by hiring alone, because the openings are growing faster than any recruiting pipeline. Plants that treat expert knowledge as an asset to be captured and distributed will get more out of the technicians they already employ, and will turn the recruits they can find into productive members of the crew far sooner.