Despite decades of Lean and Six Sigma, the consistent availability of physical assets—the machines and tools required to produce a business’s products—remains inconsistent. Overinvestment in preventive maintenance hasn’t led to improved reliability. Maintenance as a percentage of conversion cost continues to rise, with no meaningful increase in uptime. Meanwhile, assets are getting more technologically complex, the skilled labor market is tight, and downtime is less predictable and more expensive. In short, maintenance is still largely treated as a cost to control when it could be reinvented as a way to unlock margin and competitive advantage. In the chemical sector, for example, maintenance typically represents 18–23% of conversion cost. Best-in-class organizations achieve 10–12%, as shown in Figure 1.

Reliability-centered maintenance (RCM) isn't new, but it’s often executed poorly, ignored, or not adopted to its full potential. It's a smarter, more focused approach to maintenance strategy and a critical element of broader maintenance transformation. RCM aligns maintenance efforts with asset criticality (the severity of impact if the equipment failed) and the likelihood of asset failure, serving to increase reliability, consistency, and, ultimately, uptime.

What drives differential maintenance strategy?

• Asset criticality: How likely is an asset to fail based on its history, age, and condition?
• Asset likelihood to fail: Do project management plans exist for this asset and have they been used historically?
• Existing asset project management plans: Do project management plans exist for this asset and have they been used historically?
• Asset failure predictability: Has the asset failure been predictable, and are there ways to increase predictability, such as monitoring pressure?
• Cost to repair: How costly is it to repair the asset from a monetary and time perspective?

Our six-step RCM framework

RCM rarely sticks without cross-functional collaboration and expertise. So we created a sprint-style framework. An effective RCM implementation requires an operating model that bridges operations with maintenance, tools, and resources to create a sustainable solution. Our structured, scalable approach has been used successfully to drive transformation:

1. Asset segmentation

Assets are classified by their critical importance to the business. We create differentiated strategies based on the type of asset risk, such as safety, environmental, and business.

2. Criticality assessment

We evaluate the criticality and failure frequency by asset. We place special emphasis on bad actors, or assets with both high criticality and failure rates (see Figure 2).

3. Failure modes and effects analysis (FMEA) with subject-matter experts

We collaborate with subject-matter experts to identify failure modes and develop them into actionable maintenance plans.

4. Task mapping

Task mapping turns analysis into action by translating each important failure mode into a minimum set of maintenance tasks. It specifies exactly what to do, how to do it, at what frequency, and who owns it.

5. Predictive maintenance enablement

Layering predictive analytics to anticipate failures replaces manual work by a project manager. Predictive maintenance uses advanced sensors, supporting a shift toward on-demand maintenance.

6. Value validation

We compare a bottom-up costs with reliability-driven benefits to demonstrate the ROI of the new maintenance strategy.

Infusing AI into maintenance strategy

AI and large language models (LLMs) are creating new opportunities to transform maintenance by improving data quality, insights, and technician support. Companies are using AI to improve the quality of maintenance execution and to directly link maintenance activities to business value. For example, voice documentation has replaced handwritten inspection notes, allowing inspectors and maintenance operators to take more detailed notes by recording them during inspections. We projected a 35–45% reduction in the time to open a maintenance note, translating to a 1–2 percentage point increase in hands-on (“tool-in-hand”) time. AI tools then automatically convert the inspection notes describing exactly what was observed (assembly, part, how to fix it, and urgency) into the necessary format to create a service order. In areas where connectivity is limited, such as industrial plants, audio is recorded offline and automatically uploaded when connectivity is available.

In another example, we collaborated with a client to develop a technician copilot (an AI-enabled chatbot) expected to increase tool-in-hand time by 25–35% by improving planning quality. The copilot provides seamless access to information contained across multiple manuals, complex technical procedures, and various service orders. Industrial maintenance manuals are extremely long and cumbersome to use, and data retrieval is time-intensive due to poor information indexing. By providing easy access to critical information, the technician copilot helps the operator to improve safety, save time, and reduce unplanned maintenance costs.

Organizations are also using LLMs to accelerate data-intensive manual tasks that would otherwise require significant human intervention to make the data usable. The use of AI can keep the maintenance strategy evergreen and can quantify the link between equipment downtime and lost production. In one case, an energy company wanted to exclude operator comments about weather-related downtime and focus only on downtime due to equipment failures. We helped them apply an LLM to parse operator notes, identify barrels lost from non-weather-related failures, and link each event to the appropriate subsystem. This produced a clean dataset as well as an attribution for every downtime event, enabling deeper performance analysis. Most notably, this approach reduced the time required for analysis from one week to one day, freeing up operators to work on other high-value tasks.

Maintenance as a strategic lever

Reliability-centered maintenance shifts the focus from blanket prevention to targeted interventions based on asset criticality and failure likelihood. It starts with viewing maintenance as more than just a sunk cost. With a solid foundation and AI enablement, maintenance becomes a source of uptime, a path to improved margin, and a source of competitive differentiation.