The energy sector is currently navigating a period of rapid technological transition, yet many power plants remain tethered to the fragmented data practices of the 20th century. As utility operators move to repower aging facilities or integrate complex new technologies, they face a silent crisis: their most critical operational intelligence is often buried in a graveyard of disconnected EAM (Enterprise Asset Management) systems, legacy digital drives, and the fading memories of retiring subject matter experts.
A new wave of artificial intelligence (AI) integration, spearheaded by firms like 1898 & Co., is beginning to solve this "stranded data" problem. By utilizing advanced vectorization and generative AI tools, utilities are finally digitizing decades of institutional knowledge, turning chaos into a unified "source of truth."
The Core Challenge: Data Fragmentation in the Power Sector
For decades, the power generation industry has operated on a decentralized model of documentation. Information regarding piping, instrumentation, maintenance schedules, and subsystem narratives has been siloed across departments. In many cases, documentation created during the initial design and construction phases of a plant never successfully transitioned into the digital maintenance tools used today.
"This is a common challenge that we’re seeing in industry," says Chris Wiles, principal AI solutions architect at 1898 & Co. "Whether it’s to repower an old coal plant or something brand new, the data simply didn’t get encapsulated from planning to design to construction. Quite often, the most important operational intelligence is locked away in a person’s head."
When a plant undergoes a repower—the process of replacing or upgrading major components to improve efficiency or change fuel sources—this fragmentation becomes a significant risk factor. Without a cohesive data history, operators struggle to reconcile the "as-built" reality of the plant with the "as-maintained" reality found in their digital databases.
Chronology of the Transformation: From Workshops to Vectorized Intelligence
The process of reclaiming this data is not merely an IT upgrade; it is a fundamental shift in how operational knowledge is captured. The methodology developed by 1898 & Co. follows a structured evolution:
1. The Functional Audit
The process begins with a functional workshop. Rather than jumping straight into software, consultants work with plant staff to identify where information gaps exist. This phase acknowledges the reality of the "workaround culture," where employees rely on informal networks of colleagues to find documentation that should be readily accessible.
2. Centralization through Vectorization
Once the gaps are identified, the next step is to unify disparate data types. By "vectorizing" files—converting documents, diagrams, and notes into a format that AI models can interpret contextually—utilities can create a unified search environment. This allows staff to query the entire library of plant records at once, regardless of whether the source is a 30-year-old PDF, a CAD drawing, or a modern digital maintenance log.
3. Generating Institutional Narratives
The final stage involves bridging the historical gap. For many plants, the "story" of the facility—how it was built, what was changed during past overhauls, and the current state of subsystems—is missing. AI tools are now being used to generate coherent system narratives that consolidate these changes, providing a comprehensive history for new hires and maintenance teams alike.
Supporting Data: The ROI of Digital Consolidation
The practical benefits of this AI-driven approach are best demonstrated by the success seen at Cooperative Energy. Faced with the daunting task of managing 16,000 assets, the utility utilized AI to ingest data from three separate plant drives and reconcile conflicting preventive maintenance (PM) recommendations.
The $3 Million Efficiency Gain
The results were immediate and quantifiable. By automating the population of asset attributes and specifications, Cooperative Energy saved an estimated $3 million in unadjusted labor costs. This reduction in overhead was not achieved through layoffs, but by reallocating the workforce toward high-value technical tasks rather than administrative data hunting.
Rescuing "Paper-Based" Knowledge
Perhaps the most striking case involved an employee nearing retirement. The individual possessed decades of handwritten notes regarding lubrication requirements for various assets—data that had never been digitized. By scanning these notes and feeding them into the AI system, the team was able to:
- Identify conflicts between the handwritten notes and the existing EAM database.
- Verify which data was accurate through cross-referencing.
- Integrate the validated knowledge into the official record before the employee retired.
This capability effectively "future-proofs" a plant against the impending "brain drain" as the senior workforce begins to exit the industry.
Official Perspective: The Human-AI Partnership
Chris Wiles emphasizes that the primary goal of this AI implementation is to augment the human worker, not replace them. In the context of the Cleco pilot in Louisiana, the objective is to create a digital environment where staff can reliably locate their "sources of truth."
"We’re building out technology pretty quickly that manifestly changes the way that you operate a plant for the better," Wiles states. He projects that by automating the data-retrieval process, the average worker can save between one to three hours per week. In a large-scale power facility, these hours represent a massive increase in operational efficiency, allowing maintenance crews to spend more time on proactive repairs and less time searching for manuals.
Wiles argues that the capabilities of these AI models have evolved exponentially in just the last 24 months. What was once considered an impossible dream—sorting through decades of fragmented, unstructured data—is now a standard feature of modern asset management strategies.
Implications for the Industry
The shift toward AI-enabled asset management has three profound implications for the power generation sector:
1. Risk Mitigation in Repowers
As the industry faces pressure to transition toward cleaner energy, many operators are choosing to repower existing sites. A robust, AI-cleansed data foundation is essential to ensure that retrofitted components integrate seamlessly with existing infrastructure. Failure to account for legacy data often leads to costly construction delays and unexpected operational conflicts.
2. Knowledge Retention
The "Silver Tsunami"—the mass retirement of experienced engineers and plant operators—poses a significant risk to the reliability of the grid. By digitizing the "head knowledge" of these individuals through AI ingestion, utilities can capture a lifetime of expertise, ensuring that the next generation of operators inherits a structured, searchable knowledge base rather than a stack of dusty folders.
3. Operational Agility
In a competitive energy market, the speed of information retrieval is a competitive advantage. When an unexpected outage occurs, the ability to query an AI system and immediately receive accurate, verified documentation regarding a subsystem’s history can reduce downtime significantly.
Conclusion and Future Outlook
The pilot program at Cleco, and the success of the Cooperative Energy project, serve as a blueprint for the future of power generation. As utilities continue to grapple with the digital transformation, the focus must remain on bridging the gap between legacy reality and future potential.
For those interested in the technical specifics of this transition, 1898 & Co. will present a detailed case study, “AI case study: Improving asset data while setting the foundation for more efficient plant operations,” at the upcoming POWERGEN 2027 conference. The session, scheduled for Tuesday, January 19, at 10:15 a.m. in Salt Lake City, will offer further insights into how utilities can navigate the complex intersection of AI and heavy industrial infrastructure.
As the industry moves toward 2027 and beyond, the message from the field is clear: the data is there—it has simply been waiting for the right tools to bring it into the light. By investing in these AI foundations today, power plant operators are ensuring that their assets remain reliable, efficient, and ready to meet the challenges of the energy transition.
