Only 17% of manufacturing technology leaders say their organizations have fully documented, digitized and standardized their critical operational knowledge. That figure comes from new survey data from Octave, a knowledge-management specialist, and it frames a widening exposure for plant operations as experienced employees retire and AI tools spread across production floors.

The remaining 83% are split into two camps. Another 48% of respondents said core processes are documented, but the nuanced know-how accumulated by experienced employees is not. Nearly 32% said critical knowledge still moves informally, through verbal instruction and employee shadowing.

That transfer gap sits directly in the path of two converging pressures. The first is demographic: retiring veterans take undocumented troubleshooting instincts with them. The second is technological. Manufacturers are introducing AI and other digital technologies into production faster than they are preparing the workforce to use them — and the survey respondents themselves see the risk.

Approximately 38% of manufacturing respondents identified having technology but failing to upskill employees to use it as one of their greatest risks over the next five years. It ranked just behind the 39% who cited adopting AI too quickly and exposing their organizations to errors or the loss of human intuition. In other words, the top two named risks are both workforce questions, not algorithm questions.

The findings point to a specific profile for the industrial worker of the next decade: a hybrid of traditional operational expertise and broader digital fluency rather than deep specialization in data science alone.

The skill rankings back that up. Across the full industrial sample, respondents were asked which areas of study will be most valuable during the next decade. Systems thinking ranked first at 36%. Cybersecurity and communication followed at 35% each. Human-machine interaction and traditional trade or technical skills each received 34%, while engineering or computer science received 32%. Data science ranked last at 27% — a notable result for an industry being told that data talent is the bottleneck.

Octave commissioned the research firm Censuswide to conduct the survey among 756 technology decision-makers at companies with more than $1 billion in revenue, spanning manufacturing and other sectors across more than a dozen countries.

As with any vendor-sponsored research, the numbers warrant context. Octave sells knowledge-capture tooling, so it has a commercial interest in a finding that 83% of manufacturers have incomplete documentation. But the direction of the data matches what plant managers already report anecdotally: process documentation captures the procedure, not the judgment.

What to watch next: whether OEMs and tier-one suppliers begin treating knowledge capture as a formal part of launch readiness and succession planning rather than an HR afterthought, and whether the first measurable losses — quality escapes or extended downtime traced to departed expertise — force the issue before the retirement wave does.