Prescriptive Analytics: The Holy Grail of Condition-Based Maintenance
For over sixty years, machinery-intensive industries have relied on some form of condition monitoring. Yet for most of that time, the cost and complexity of implementation limited its use to the most critical assets, and organisations remained heavily reliant on time-based maintenance.
That era is over. The introduction of wireless sensors and purpose-built AI has removed barriers to adoption, and condition monitoring has transitioned from a "nice to have" to a "must have" operational necessity. Moreover, it is no longer enough to simply know if a machine might fail. Faced with resource shortages and cost pressures, organisations want to know why a machine is failing and exactly how to fix it before there’s an impact on operations.
This is why organizations are turning to prescriptive analytics, the holy grail of efficient condition-based maintenance.
In this article, examine the benefits of prescriptive analytics and how advances in AI support broad adoption and accelerate time to value.
Prescriptive analytics: Moving from insight to action
The industry is currently pivoting from simple prediction to prescription.
While predictive analytics asks, "When will this fail?", prescriptive analytics asks, "What is the root cause, and how do we fix it?". AI is enabling this shift. AI can process data at a speed and scale that human analysts simply cannot. It can also be paired with physics-based models for more efficient and accurate diagnostics.
We have introduced prescriptive analytics with Cordant™ Machine Health, which detects, diagnoses, and prescribes fixes for machine issues by analyzing millions of condition monitoring sensor outputs 24/7.
The power of purpose-built AI: Speed to value
Integrating general-purpose AI into asset management can be complex, often taking months or longer to deploy and "learn" your machinery before returning value.
The alternative is a purpose-built AI solution like Cordant™ Machine Health powered by Augury. The solution combines Bently Nevada’s sensing capabilities and hardware with Augury’s cutting-edge AI-driven machine health technology to deliver prescriptive analytics. With Augury’s models trained on a massive library of failure modes and over 3 million hours of monitoring data, they can also deliver specific diagnostic insights almost immediately.
This allows operators to skip the long deployment phase and instantly begin to analyze and diagnose issues. As Cordant™ Machine Health validates data against physics-based models, it can also rule out false positives—a common issue where up to 70% of alerts can be false alarms.
Teams can then focus on the alerts that matter and take proactive action before issues escalate, effectively moving maintenance from a time-based guess to a condition-based science.
What to consider when adopting a prescriptive analytics solution
There are several practical considerations when evolving your condition-based maintenance approach and adopting a prescriptive analytics solution.
1. There is no "one size fits all" solution. Production assets may require different monitoring strategies based on their criticality to production, failure modes, operating environment, and redundancy. It is essential to audit assets to identify the right monitoring solution for each type.
2. Siloed data is a barrier to efficiency. Working with a single vendor or a unified platform ensures that systems are designed to work together. When you have this level of integration, you avoid a lot of gaps in asset coverage, in data collection, and in communication between systems and teams in the event of an emergency. One call for one fix is a far better strategy than multiple calls for one fix, with multiple different vendors approaching a problem from various siloed, and limited, points of view.
3. Evaluate "As-a-Service" Models. When in-house resources are scarce, "as-a-service" models offer a lifeline, providing access to specific capabilities, expert support, and software without heavy upfront infrastructure costs. This allows operators to access expert intelligence, including vibration analysts and reliability engineers, for an additional layer of support.
Real-world impact: Solugen moves from reactive to prescriptive maintenance and avoids more than 100 hours in downtime.
The value of prescriptive analytics can be demonstrated by our work with the chemicals company Solugen.
Solugen partnered with Baker Hughes’ Customer Success team to deploy machine health capabilities from Cordant™ Asset Health. The solution's simple architecture—consisting of a sensor, gateway, and cloud-based platform for monitoring and analytics—allowed for quick and scalable implementation.
In the first two years of implementation, Cordant™ has helped Solugen avoid 122 hours of downtime and achieve over $111.3K in cost savings. These savings represent a 2x return on investment and include avoidance of costs associated with downtime, labor, parts, and waste.
The future is prescriptive
As industries face increasing demands to cut costs and improve reliability, the old ways of operating are no longer sufficient.
Prescriptive analytics represent the maturity of condition monitoring. It moves us beyond manual data collection and false alerts toward a future where machines tell us not just if they are sick, but how to heal them. By embracing data-driven strategies and leveraging purpose-built AI, organizations can minimize unplanned downtime and maximize operational efficiency.
Learn more: