Key Takeaways
- Critical operational knowledge often resides with a small number of experienced engineers and planners.
- Much of what organizations know about their assets, systems, and maintenance practices is not fully documented.
- Stale job plans and undocumented changes make it more difficult to capture and transfer expertise.
- Knowledge graphs and AI can help preserve institutional knowledge before it is lost.
The People Who Know How This Equipment Fails Are Leaving
A meaningful portion of the energy and utility maintenance workforce is retirement-eligible within the next five to eight years at many large organizations. This is documented and widely acknowledged. Most organizations have not built the infrastructure to capture the knowledge that will leave with those engineers.
The knowledge at risk is specific. It is the field engineer who knows from twenty years of experience which failure mode on a particular asset design typically appears three months before the unit fails, and what to look for. It is the reliability engineer who wrote the original FMEA tables for a facility’s major equipment and holds in their head the reasoning behind the maintenance intervals that now exist only as numbers in the system. It is the senior planner who knows which job plans in the CMMS are trustworthy and which ones have not been updated since the last major system migration.
But this knowledge loss is not necessarily tied to a future retirement cliff. My dad, who worked more than 40 years at a major utility company and is long retired, still gets calls about ancient substations built in the early 80s that he helped design. This is happening today.
"The brain drain is an ongoing battle everywhere across the industry.”
The Gap Between What Is Known and What Is Documented
In many cases, organizations are also managing differences between what was originally documented and what actually exists in the field today. Equipment is replaced. Infrastructure changes. Systems evolve.
Maintenance activities are often based on assumptions about how equipment, infrastructure, and systems interact. There can be areas of a facility that no one fully understands until something breaks. At an electric utility, you can schedule maintenance on a circuit, bring that circuit down, and then discover knock-on effects you did not know about.
Job plan quality is a concrete measure of this problem. Industry estimates consistently put 30 to 60 percent of active CMMS job plans as either missing or last updated more than three years ago.
When a job plan is stale or absent, technicians improvise. When they improvise, as-found and as-left conditions may go uncaptured. When those conditions go uncaptured, the failure data does not make it back into the system in a form that informs future maintenance decisions.
What should be a mechanism for capturing institutional knowledge instead becomes another point where knowledge is lost.
Capturing Knowledge Before It Walks Out the Door
The best time to address this is right now.
Organizations can work directly with experienced engineers to document changes that were never formally captured. Those changes often include equipment modifications, maintenance practices, operational workarounds, and decisions that exist only in individual experience.
That knowledge can then be connected through an AI-powered knowledge graph alongside maintenance records, historian data, sensor data, and operational systems to create a more complete picture of how systems actually operate. By mapping relationships between assets, failure modes, interventions, and outcomes, knowledge graphs make expertise easier to search, transfer, and apply.
AI can also synthesize work order history, OEM documentation, and maintenance records into draft job plans that a credentialed planner reviews and approves, while Generative AI interfaces can give field technicians access to relevant expertise and operational context at the point of work.
This Challenge Gets Compounded
The risk of knowledge loss is compounded by the issues discussed in my previous article on maintenance orchestration: poorly structured data, manual coordination across disconnected systems, and work orders that lack the context needed to support efficient execution.
When institutional knowledge goes uncaptured, maintenance strategies become harder to improve. Job plans become harder to maintain. Stale job plans accelerate the knowledge retirement problem because the procedures that should be capturing institutional knowledge are themselves out of date.
And when that knowledge is lost, the quality of the data feeding future maintenance decisions deteriorates over time. The gap is in deploying these tools before the knowledge retires, not after.
Perficient works with energy and utility organizations, including oil and gas operators, across the full maintenance orchestration stack. Explore more of our Automotive & Industrials expertise.