Generative AI benefits for asset lifecycle management

Generative AI insights

• By expanding from analysis and prediction to creation, generative AI opens new use cases for asset lifecycle management (ALM) like faster failure mode effect analysis (FMEAs), conversational maintenance assistants and more.
• To capture value from generative AI, businesses need to embrace a second wave of adoption by identifying process inefficiencies, gaps in data and connecting the right stakeholders throughout an organization.

Over the past few years, artificial intelligence (AI)  has been driving incredible progress in predictive maintenance, enabling smarter, real-time monitoring and improving the uptime and efficiency of assets. With the emergence of generative AI, even more areas of asset lifecycle management (ALM) are taking leaps ahead and unlocking even more options to minimize maintenance costs, optimize physical assets, and contribute to sustainability and energy cost goals—even for systems that were, until recently, considered optimized.

The evolution of AI for ALM processes

Businesses are already well into the heady days of AI. Many collect data on almost every aspect of their assets: Operational data from their facilities; resource data on electricity and other utilities; asset data for tracking age, condition, maintenance history; and replacement parts for machines in factories. Many have brought together various databases and metrics under one dashboard, cleaned data of outliers or errors, labeled data and even started tuning or training AI models for their specific needs.

Collecting this data and leveraging AI-driven insights is important, and has already had a real impact, giving some enterprises an advantage that helps differentiate them as leaders in front of the pack.

Figure 1: IBM Maximo Work Order Intelligence uses generative AI to recommend possible failure codes, with a description and confidence level for each option. Courtesy: IBM

For example, computer vision—where machine learning helps identify and understand people and objects—helped Ford Motor Company reduce defects on the assembly line, from 40 per month to zero. The zero-defect system even helped drive design improvements.

In another example, computer vision is analyzing photos from drones and other sources to help optimize maintenance for the Great Belt bridge in Denmark, which is projected to extend the bridge’s life by 100 years.

In some ways, generative AI continues the trend of technological optimization, but in other ways it is another step-change forward. Because generative AI by nature can create—in addition to analyze, extract, predict and more—it allows businesses to go one step further and tackle new use cases specific to ALM.

In the examples above, generative AI would help clients move beyond the object recognition stage—already hugely beneficial to businesses—and engage in more sophisticated communication with people throughout the business.

Narrow the skills gap with data from generative AI

One foundational benefit of generative AI is how it helps to bridge the skills gap that many employers are facing in today’s competitive environment. For example, instead of requiring a field technician to be prepared for the huge variety of issues they may face at any point, conversational chat models are helping quickly surface the right information at the right time. When facing a problem, a field tech can use natural language to learn more about a specific type of asset or a recent asset failure. He or she can input data one piece at a time and let the AI do the heavy analysis, which might include asking the tech for more data to help find with the root cause of an issue with a speedy analysis.

Generative AI also is “under the hood” when employees or companies are practicing ALM. Many of the insights driven by generative AI rely on time series forecasting models, which can analyze historical data on how assets have changed over time, identify anomalies and make predictions as to what to expect for a given asset. These AI-generated predictions can also help enterprises avoid expensive and unnecessary shutdowns.

Figure 2: IBM Maximo Assistant uses conversational gen AI (right side) to help a user create a work order. Courtesy: IBM

In similar ways, generative AI can help speed completion of failure mode effect analysis (FMEA). When asset failures happen—as they are bound to, especially in contexts like manufacturing—it is critical for companies to be as prepared as possible with the necessary parts and expertise on standby. Otherwise, unexpected failures can lead to extraordinary costs in downtime and dollars. Therefore, it’s critical to not only quickly pinpoint the cause of asset failure but also make the necessary changes to avoid repetition. Completing a step-by-step FMEA identifies all possible causes of failure is time consuming.

Generative AI can streamline the process in several ways. For example, virtual assistants can provide all the possible failure causes quicker than a single (or group) of employees can, then help narrow down the most likely origin of the failure, and contribute to reaching a solution in a more timely manner. This use case is already being implemented with software offerings that provide businesses libraries full of asset-specific failure details and mitigation activities so they can proactively plan to anticipate failures, and avert or quickly remediate them.

Figure 3: IBM Maximo Operational Dashboard offers an AI-generated insight on why an asset may be suffering faulty connections. Courtesy: IBM

Research also shows training AI models on a library of asset reliability strategies can generate a tremendous amount of key content to build FMEAs. In the future, generative AI will be capable of compiling this content and provide organizations with custom suggestions on when and how to maintain their assets to avoid these failures.

Generative AI can also help mitigate data challenges. For many organizations, a key challenge to moving forward with AI is they are worried about their data quality; they don’t want to create about a “garbage in, garbage out” scenario.

Generative AI can help in several ways. For example, layering generative AI on top of other diagnostic AI tools can help improve the accuracy of the “failure codes” assigned when ALM problems occur. Those codes often are missing or incorrect, but now tools can automatically generate failure code recommendations by training models on long and short descriptions from relevant work orders.

What’s next for generative AI in ALM

With technology evolving faster than ever, businesses don’t have the luxury of resting on their opening advances with AI. Generative AI is enhancing employee skills, disrupting the talent landscape, reducing the time needed for complex diagnoses, and helping to extract more insights from less data. There’s still opportunities for organizations to get “ahead of the curve.”

Figure 4: IBM Maximo Visual Inspection uses AI computer vision to identify and label anomalies. Courtesy: IBM

Whether companies are looking to improve decision making, asset performance, workforce productivity or all of the above, there are ways to leverage generative AI. First, identify the gaps in the current processes. Second, identify what data is needed, tap into the right colleagues and experts and start to collect what’s available.

Imagine how generative AI might be able to help fill the remaining gaps. We have already seen this happen with the first wave of AI adoption; it is time to move forward with the second.

Mike Hollinger, chief architect, IBM Maximo AI Software and Solutions. Edited by Chris Vavra, senior editor, Control Engineering, WTWH Media LLC, cvavra@wtwhmedia.com.

Do you have experience and expertise with the topics mentioned in this content? You should consider contributing to our WTWH Media editorial team and getting the recognition you and your company deserve. Click here to start this process.